Upload
vanminh
View
213
Download
0
Embed Size (px)
Citation preview
i
Generation, stability and migration of montmorillonite colloids in aqueous systems
Sandra García García
Doctoral Thesis
School of Chemical Science and Engineering
Royal Institute of Technology
Stockholm, Sweden 2010
AKADEMISK AVHANDLING
Som med tillstånd av Kungliga Tekniska Högskolan framlägges till offentlig granskning
för avläggande av Doktorsexamen i kemi fredagen den 29 januari 2010, kl. 10.00 i sal F1,
Kungliga Tekniska Högskolan, Lindstedtsvägen 22, KTH Stockholm
Opponent: Dr.Tiziana Missana
ii
ISBN: 978-91-7415-535-8
ISSN 1654-1081
TRITA-CHE Report 2010:3
© Sandra García García 2010
Tryck: Universitetsservice AB, Stockholm 2010
iii
To my family
iv
v
Abstract
In Sweden the encapsulated nuclear waste will be surrounded by compacted bentonite in
the granitic host rock. In contact with water-bearing fractures the bentonite barrier may
release montmorillonite colloids that may be further transported in groundwater. If large
amounts of material are eroded from the barrier, the buffer functionality can be
compromised. Furthermore, in the scenario of a leaking canister, strongly sorbing
radionuclides, can be transported by montmorillonite colloids towards the biosphere. This
thesis addresses the effects of groundwater chemistry on the generation, stability,
sorption and transport of montmorillonite colloids in water bearing rock fractures.
To be able to predict quantities of montmorillonite colloids released from the bentonite
barrier in contact with groundwater of varying salinity, generation and sedimentation test
were performed. The aim is first to gain understanding on the processes involved in
colloid generation from the bentonite barrier. Secondly it is to test if concentration
gradients of montmorillonite colloids outside the barrier determined by simple
sedimentation experiments are comparable to generation tests. Identical final
concentrations and colloid size distributions were achieved in both types of tests.
Colloid stability is strongly correlated to the groundwater chemistry. The impact of pH,
ionic strength and temperature was studied. Aggregation kinetics experiments revealed
that for colloid aggregation rate increased with increasing ionic strength. The aggregation
rate decreased with increasing pH. The temperature effect on montmorillonite colloid
stability is pH-dependent. At pH≤4, the rate constant for colloid aggregation increased
with increasing temperature, regardless of ionic strength. At pH≥10, the aggregation rate
constant decreased with increasing temperature. In the intermediate pH interval, the
aggregation rate constant decreased with increasing temperature except at the highest
ionic strength, where it increased. The relationship between the rate constant and the
ionic strength allowed the critical coagulation concentration (CCC) for Na- and Ca-
montmorillonite to be determined.
In order to distinguish the contribution of physical filtration and sorption to colloid
retention in transport, the different retention mechanisms were quantified. Sorption on
different representative minerals in granite fractures was measured for latex colloids (50,
100, 200 nm) and montmorillonite colloids as a function of ionic strength and pH.
Despite of the negative charge in mineral surfaces and colloids, sorption was detected.
The sorption is correlated to the mineral point of zero charge and the zeta potential of the
colloids, and increases with increasing ionic strength and decreasing pH. In transport
experiments with latex colloids in columns packed with fracture filling material, the
retention by sorption could clearly be seen. In particular at low flow rates, when the
contact time for colloids with the mineral surfaces were the longest, sorption contributed
to retention of the transport significantly. The retention of latex colloids appeared to be
irreversible in contrary to the reversible montmorillonite colloid retention.
Generation, stability and sorption of the montmorillonite colloids are controlled by
electrostatic forces; hence, the results were in qualitative agreement with DLVO.
vi
vii
Sammanfattning
I Sverige kommer kapseln som innehåller det urbrända kärnbränslet att omges med
kompakterad bentonitlera. Vi kontakt med vattenförande spricka kan bentonit avge
kolloidala partiklar som kan transporteras med grundvattnet. Om stora mängder material
eroderas bort från barriären äventyras dess funktion. I händelse av en läckande kapsel kan
lerkolloider bidra till spridning av starkt adsorberande radionuklider. I denna avhandling
studeras effekter av grundvattenkemi på generering, stabilitet, adsorption och transport av
montmorillonitkolloider i vattenbärande bergsprickor.
För att kunna förutsäga mängden monmorillonitkolloider som frisätts av
bentonitbarriären i kontakt med grundvatten av varierande salthalt har försök med
avseende om generering och sedimentation utförts. Dessa försök visade att den slutgiltiga
koncentrationen av kolloider är den samma i båda typerna av försök.
Kolloidstabilitet är starkt knutet till grundvattenkemi. Effekten av pH, jonstyrka och
temperatur har därför studerats. Dessa försök visade att hastigheten för
kolloidaggregering ökade med ökande jonstyrka och minskade med ökande pH.
Temperatureffekten är pH-beroende. Vid pH lägre än 4 ökar hastighetskonstanten för
aggregering med ökande temperatur, oberoende av jonstyrka. Vid pH högre än 10
minskar hastighetskonstanten med ökande temperatur. I pH intervallet 4-10 minskar
hastighetskonstanten med ökande temperatur utom vid höga jonstyrkor där trenden är den
motsatta. Sambandet mellan kinetik för aggregering och lösningens jonstyrka
möjliggjorde bestämning av den kritiska koaguleringskoncentrationen (CCC) för både
Na- och Ca-montmorillonit.
För att kunna skilja mellan retention pga fysikalisk filtrering och adsorption vid
kolloidtransport studerades dessa processer separat. Adsorption av latexkolloider (50, 100
och 200 nm) och montmorillonitkolloider på representativa mineralytor studerades som
funktion av jonstyrka och pH. Trots att både kolloider och mineralytor var negativt
laddade under rådande betingelser observerades adsorption. Affiniteten för kolloider till
mineralytor visade sig bero på nolladdningspunkten (pzc) för mineralet och
zetapotentialen för kolloiden. Affiniteten ökar med ökande jonstyrka och sjunkande pH. I
transportförsök med latexkolloider i kolonner packade med sprickfyllnadsmaterial kunde
retention pga sorption tydligt ses. Detta var tydligast vid låga flöden där uppehållstiden i
kolonnen är lång. Retentionen av latexkolloider var irreversibel i motsats till retentionen
av montmorillonitkolloider som visade tecken på reversibilitet.
Samtliga resultat är i god överensstämmelse med DLVO-teorin.
viii
ix
List of Publications
This doctoral thesis is based on the following papers and manuscripts, which are referred
to in the text by their Roman numerals:
I. Determining pseudo-equilibrium of montmorillonite colloids in generation
and sedimentation experiments as a function of ionic strength, cationic form,
and elevation
S. García-García, C. Degueldre, S. Wold, S. Frick
Journal of Colloid and Interface Science, 335 (2009) 54–61
II. Kinetic determination of Critical Coagulation Concentrations for Sodium-
and Calcium-Montmorillonite colloids in NaCl and CaCl2 aqueous solutions.
S. García-García, S. Wold, M. Jonsson
Journal of Colloid and Interface Science, 315 (2007) 512-519
III. Temperature effect on the stability of bentonite colloids in water.
S. García-García, M. Jonsson, S. Wold
Journal of Colloid and Interface Science, 298 (2006) 694-705
IV. Effect of pH and temperature on the stability of bentonite colloids.
S. García-García, S. Wold and M. Jonsson
Applied Clay Science, 43 (2009) 21-26
V. Colloid sorption on minerals. Effects of colloidal size, pH, buffer
concentration, and mineral properties
S. García-García, S. Wold and M. Jonsson
Submitted to Applied Geochemistry
VI. Colloid transport in fracture filling materials
S. García-García, S. Wold and M. Jonsson
Submitted to Journal of Contaminant Hydrology
Comment on my contribution to the publications
Paper I, I contributed to the design of the experiments and performed most of the
experiments. I have participated in the analysis of the results and I wrote the manuscript
in collaboration with the co-authors.
Paper II III IV V and VI, I have designed and performed the experiments. I have
carried on the simulations and participated in the analysis of the results and I wrote the
manuscript in collaboration with the co-authors.
x
xi
Table of Contents
Generation, stability and migration of montmorillonite colloids in aqueous systems ......... i 1. Introduction ..................................................................................................................... 1
1.1. Background of the project ........................................................................................ 1 1.2. Scope of the thesis ................................................................................................... 3
1.3. The bentonite barrier ................................................................................................ 4 1.4. The mineral montmorillonite ................................................................................... 6 1.5. Colloids .................................................................................................................... 7 1.6. Generation, aggregation and sedimentation of colloidal particles ........................... 7 1.7. The electrical double layer ....................................................................................... 9
1.8. Total inter-particle energy...................................................................................... 10 1.10. Colloid retention processes in water bearing fractures ........................................ 13
2. Materials ....................................................................................................................... 15 2.1. Bentonite ................................................................................................................ 15
2.2. Sodium- and calcium-montmorillonite .................................................................. 16 2.3. Reference colloids .................................................................................................. 17
3. Methods......................................................................................................................... 21 3.1. Single Particle Counting ........................................................................................ 21 3.2. Photon Correlation Spectroscopy .......................................................................... 21
3.3. Zeta potential analysis............................................................................................ 23 3.4. Fluorescence spectrophotometry ........................................................................... 23
4. Results and Discussion ................................................................................................. 25 4.1. Colloid generation .................................................................................................. 25
4.2. Stability of colloidal suspensions........................................................................... 29 4.2.1. Method applicability and PCS calibration .................................................. 29
4.2.2. Effect of the ionic strength on aggregation kinetics ................................... 31 4.2.3. Influence of temperature on aggregation kinetics ....................................... 37 4.2.4. Ionic strength, pH and temperature effects on the stability of
montmorillonite colloids ....................................................................................... 41
4.3. Adsorption of colloids on mineral surfaces ........................................................... 45 4.4. Colloid transport .................................................................................................... 47
5. Concluding remarks ...................................................................................................... 53 6. Acknowledgements ....................................................................................................... 54 7. Bibliography ................................................................................................................. 55
xii
1
1. Introduction
1.1. Background of the project
Colloidal particles suspended in deep bedrock groundwaters are generated as the result of
a number of processes such as mineral weathering, precipitation, erosion, biological
activity and organic matter degradation. The suspended particles have diameters ranging
from nanometers to micrometers, and varying chemical compositions.1-4
The small size
of colloidal particles gives a large surface-to-volume ratio as well as lesser effect of the
gravitational force. In addition, the surface groups of colloids often give raise to surface
charge in natural waters. These properties allow colloidal particles to remain in
suspension, to diffuse by Brownian motion and to sorb contaminants from the
groundwater by electrostatic interactions, surface complexation, precipitation,
polymerization or ion exchange.5 Association of contaminants with mobile colloidal
particles, is an important environmental issue since contaminants are often spread with
colloids in nature.6,7
Hence, understanding the mechanisms of generation, stability and
mobility of colloidal particles in nature is of great importance for predicting the fate of
contaminants in the environment.
One particular case where colloid formation and migration is of crucial importance is
deep geological repositories for highly radioactive nuclear waste. Deep geological
repositories for spent nuclear fuel are constructed to isolate the radioactive waste from
the biosphere for a period of time as long as 100.000 years. Many countries carry out
active research programs in order to take care of their own nuclear waste. The main
repository alternatives are natural clay and crystalline rock formations. Clay formations
have been adopted by France and Belgium,8,9
while Finland and Sweden will build their
repositories in crystalline rock.10,11
The company Posiva Oy in Finland has already
started to excavate the tunnel at the Olkiluoto site, and The Swedish Nuclear Fuel and
Waste Management Co (SKB) in Sweden has recently suggested Forsmark as the most
suitable location for the repository. The disposal facilities in Sweden and Finland will
both be constructed in granite bedrock in accordance with the KBS-3 concept developed
by SKB.
The Swedish KBS-3 concept will fulfill the safety requirements by a multi-barrier system
consisting of four engineered and natural barriers12
as represented in Figure 1. In a multi-
barrier system, the different barriers support and complement each other but they are
intended to be functionally independent in the event of the possible failure of any of the
other barriers:13,14
- The spent fuel itself constitutes the first barrier. The spent fuel consists of 95% UO2 and
5% fission products and actinides. In the event of groundwater intrusion due to barrier
failure, radionuclides will be released at the rate of UO2 (fuel matrix) dissolution.
Uranium dioxide has low solubility under reducing conditions.15
2
- The second barrier is a copper canister provided with a cast iron insert in which the fuel
elements will be encapsulated. The copper canister is does not corrode under reducing
conditions.16
- The third barrier is the compacted saturated bentonite that will surround the canisters. It
will provide mechanical support for the canisters. Another beneficial feature is that the
plasticity of the swollen bentonite will minimize the potential damage caused to the
canister by any future rock movements.17
When water-saturated, the bentonite barrier has
low hydraulic permeability which will reduce water flow around the canisters. Due to the
low hydraulic permeability, the only transport mechanism through the barrier is
diffusion,18
which will limit the access of corrosive species such as HS- to the canister
and will retard escaping radionuclides. The migration of positively charged radionuclides
will be further retarded by surface complexation and cation exchange processes in the
bentonite matrix.
Figure 1: The KBS-3 concept for disposal of spent nuclear fuel.6
- The crystalline granite host rock is the fourth barrier that ensures a stable mechanical
and chemical environment to the canisters and where sorbing surfaces for radiunuclides.19
The host rock has to fulfill a number of requirements in order to be suitable for a deep
repository. One important safety requirement is to have a low fracture frequency since the
presence of groundwater is undesirable due to the risk of canister corrosion and transport
of radionuclides in groundwater.20
Of the two sites that were under consideration in
Sweden, this requirement is best fulfilled by the Forsmark site. Groundwater at the
Forsmark site is generally anaerobic with reported 0.032-0.052 mL L-1
dissolved oxygen.
It has a pH of 7.7 and a redox potential (Ag/AgCl) of –281 mV. The natural colloid
content measured by Laser-Induced Breakdown Detection (LIBD) was 30–50 μg L-1
at
3
437 m depth.21
The flow measured in 32 boreholes was found to be 0.01 to 123 mL min-1
with calculated Darcy velocities from 6.7∙10–11
to 6.6∙10–7
m s-1
.22
Although the presence of fractures will be avoided, there will be some water-bearing
fractures that will intercept the deposition holes. Under present conditions in the selected
host rock, release and transport of colloidal particles is not likely to occur since the
groundwater salinity is high. However the physicochemical conditions in permeable
fracture formations may change23,24
and thereby enhance or inhibit colloid transport.
For instance, in contact with low saline groundwaters after calcium depletion,25,26
a
highly hydrated bentonite gel may form, consisting of charged aluminosilicate sheets
weakly bound to each other by electrostactic and van der Waals forces. At the bentonite-
water interface, disintegration of the gel into colloidal particles could take place if
repulsion between charged sheets is strong enough. If bentonite colloidal particles are
released from the buffer barrier in large quantities, the functionality of the bentonite
barrier will decrease. Loss of bentonite material would imply higher hydraulic
conductivity, which in turn will allow corrosive species to reach the canisters faster,
thereby increasing the risk of corrosion and radionuclide leakage.
1.2. Scope of the thesis
The aim of this work is to understand montmorillonite colloid stability and transport in
water-bearing fractures. More specifically, this has been done by:
- Studying colloid generation from Na- and Ca-montmorillonite as a function of ionic
strength.
- Studying the effect of water chemistry on colloid stability; the impact of ionic strength,
pH and temperature on Na- and Ca-montmorillonite.
- Studying the sorption of colloids on mineral surfaces at static conditions.
- Studying the migration of montmorillonite colloids in columns packed with fracture
filling material; the role of retention processes.
A schematic illustration of the scope of this thesis is given in Figure 2.
4
Bentonite barrier
in contact with
Water-bearing fractures
1. Does Colloid generation occur?
(Paper I)
2. Are Colloids stable?
(Paper II, III, IV)
3. Do Colloids sorb onto Host Rock?
(Paper V)
4. What is the impact of retention
on Colloid transport?
(Paper VI)
DLVO
Theory
Figure 2: Schematic representation of the scope of this thesis.
1.3. The bentonite barrier
The motivation for using bentonite as a buffer material is its high content (65-90%) of the
mineral montmorillonite. Montmorillonite has pH buffering capacity and swells in
contact with water. These properties will provide a stable physical and chemical
environment for the canisters (see Figure 3). When saturated with de-ionized water to a
density of 2000±50 kg m-3
, compacted bentonite has a swelling pressure of about 104 kPa
and a hydraulic conductivity27
of 7∙10-14
m s-1
. Therefore, the water flow around the
canister will be reduced by the buffer barrier, preventing corrosive agents such as
sulphide from coming into contact with the canister.25
In addition, montmorillonite has excellent sorption characteristics. In the event of
radionuclides being released from a failed canister, the mobility of cationic contaminants
will be reduced, since they will be retarded by surface complexation and cation exchange
reactions.
During the lifetime of the repository, the fracture system, the groundwater composition
and flow may be affected by climate changes. The potential changes in the climate and
the consequences for the repository are difficult to predict.
Scandinavia is expected to experience glacial cycles, during which the ice sheet grow and
retreat and melt water accumulates and can infiltrate by pore pressure differences
generated at both sides of the ice sheet margin. If larger fractures open by the pressure
5
released after ice retreat, the melt water may with high flow rates, reach the repository,
displacing the more saline pre-existing groundwaters without mixing.23,24
In this scenario, these processes will lead to hydrological changes in permeability,
groundwater pressure, groundwater flow, groundwater salinity, pH and oxygen content at
repository depths.28
Figure 3: Deposition chamber with bentonite buffer and canister.
14
The importance of colloid release lies not only in the loss of buffer material, but also in
the risk of colloid-facilitated radionuclide transport, since in the event of canister failure
most radionuclides will adsorb strongly onto bentonite. It is known that the migration
velocity of soluble contaminants in groundwater is retarded due to matrix sorption.29
However, when sorbed to colloids, contaminants get mobilized and migrate faster.7,30
6
1.4. The mineral montmorillonite
Montmorillonite is a layered aluminosilicate mineral from the 2:1 smectite group. Each
layer consists of three sheets: an octahedral sheet between two tetrahedral sheets, as
illustrated in Figure 4. Aluminium atoms are present at the octahedral sites, coordinated
to oxygen and hydroxyl groups, while silicon atoms coordinate to oxygen in the
tetrahedral positions.
Figure 4: Na-montmorillonite structure.
Isomorphous substitution occurs mainly in octahedral sheets, where Al is replaced by Mg
or Fe, but substitution of Si by Al in the tetrahedral sheets can also take place. As a
consequence, the layers have a permanent negative charge with a charge density of about
0.13 C m-2
. The excess of negative charge is compensated for by adsorbed positive ions
such as Na+ and Ca
2+ between the layers. The binding forces between layers are much
weaker than those within sheets. When montmorillonite comes into contact with water,
exchangeable cations hydrate and successive water layers occupy the interlayer space and
the montmorillonite swells. The extent of swelling depends on the compensating cations
and available volume. Under constricted volume, the incorporation of well ordered water
layers is known as crystalline swelling. Under free-swelling conditions, more than four
water layers may be incorporated by Na-montmorillonite by osmotic swelling until
repulsive and attractive forces between sheets reach equilibrium. Ca-montmorillonite on
the other hand, shows little if any osmotic swelling.31
Once the montmorillonite layers
have been filled with water molecules, the compensating cations can easily diffuse out
and be exchanged by new ions from the solution/liquid phase. This feature is
characteristic of many clay minerals. The typical cation exchange capacity in
montmorillonite is about 0.9 meq g1.18
7
1.5. Colloids
A colloidal dispersion is a distribution of small particles of a substance (solid, liquid or
gas) in a continuous phase. The size of the particles ranges from nanometers up to
micrometers.
Given the small size (1 nm- 1 µm) of the suspended particles, they have a large surface
area per unit mass. Due to their low mass, colloidal particles are not strongly affected by
gravitational force. Most particles in suspension bear a surface charge, which results in
repulsive forces between particles, preventing them from agglomerating. Voluminous
groups binding or adsorbing on the surface can also prevent particle agglomeration by
osmotic and volume restriction effects.32
The combination of a low tendency to
agglomerate and the low effect of the gravitational force results in stable colloidal
dispersions.
Similar to ions and molecules, colloidal particles in a fluid undergo random
displacements and collisions. This phenomenon is known as Brownian motion. As a
result, colloidal particles are transported by diffusion and distribute homogeneously in a
liquid. 33
Depending on the water affinity of the dry solid, colloids are traditionally classified as
lyophilic or lyophobic, where the lyophobic colloids have lower affinity for water.
Another common classification used refers to the stabilization mechanism that prevents
particles from sticking together, i.e., steric stabilization or electrostatic stabilization.34
Depending on the size distribution, colloids can be classified as monodisperse or
polydisperse. In principle, in a monodisperse suspension all the particles have the same
shape and size, while in polydisperse suspensions the particles differ in size and shape.34
Most of the organic and inorganic colloids in natural systems have broad size
distributions.35,36,37
1.6. Generation, aggregation and sedimentation of colloidal particles
Particles of clay minerals, iron hydroxides, silica and natural organic matter among others
are present in natural aquifers in the colloidal size range and their concentrations are
affected by hydrogeochemical perturbations.38,39
Montmorillonite colloid generation can occur when the material is in contact with water.
Swelling leads to the formation of a gel-like front moving with the montmorillonite
concentration gradient. Due to high hydration of montmorillonite in the gel structure, the
attractive forces between the negative sheets and the hydrated cations become weaker and
colloidal particles start to disjoin and diffuse from the montmorillonite/water interface.
This is the mechanism for generation of montmorillonite colloidal particles. In the case of
8
Ca-montmorillonite the swelling capacity that induces gel propagation is very low
compared to Na-montmorillonite.
The swelling pressure of montmorillonite, and hence colloid generation, increases with
decreasing ionic strength, increasing compaction density and pH.40
This can be
understood by considering the effects of these parameters on the electrostatics on the
colloidal montmorillonite particles.
When two particles collide they can form an aggregate. If the particles undergo
irreversible aggregation, the colloids coagulate, while if the aggregates may be
redispersed by shaking or changing the conditions in the system, the colloids flocculate.
Flocculated lyophobic suspensions redisperse if the conditions change leading to an
increase in the stability of the particles.
In order for small particles to sediment, it is necessary to apply a centrifugal field
stronger than the normal gravitational field. Large aggregates deposit at a uniform rate
determined by the ratio between the gravitational and friction forces. The sedimentation
rate of a large aggregate due to the gravitational field can be expressed as:
fgmfgvdtdx /)/1(/)(/ 00 (1)
where g is the acceleration due to gravity in N kg−1
, m is the mass of the particle in kg, v
is the volume of the particle (m3), ρ is the density of the particle (kg m
-3), ρ0 is the density
of the solution and f is the frictional coefficient, proportional to the viscosity of the
medium η in Pa s and the radius of the particles a in metres.
41
af 6 (2)
After aggregation the probability of sedimentation increases, since both processes are
intimately connected. When aggregation is the limiting step, sedimentation of the large
aggregates is faster than with their rate of formation. Hence, in this case the size
distribution of the particles does not change significantly with time in the upper part of
the suspensions and aggregation exhibits second order kinetics (Paper II, III, IV).
Colloid aggregation can be represented by a bimolecular reaction:42
A + A → A2 (3)
For reaction (3), the rate can be expressed as:
22 Akdt
Ad (4)
where k is the rate constant for the aggregation process and [A] is the concentration of
particles. Integration of equation (4) leads to:
9
ktAA t 2/1/1 0 (5)
The slope of the line obtained when 1/[A]is plotted versus t is 2k.
The aggregation rate for particles in the diffusion-controlled regime is given by
expression (6):43
216/ AaDdtAd p (6)
where t is time, Dp is the diffusion coefficient and a is the particle radius. The diffusion
coefficient, Dp can be defined as:
aTkD Bp 6/ (7)
where kB is the Boltzmann constant, T is temperature and η is the viscosity of the
medium. Expression (6) can be rewritten as:
23
8/ A
TkdtAd B
(8)
where 3
8 TkB is the diffusion-controlled rate constant, which is independent of the radius
and nature of the particles and only depends on the viscosity of the medium. For dilute
aqueous suspensions, the diffusion-controlled rate constant for colloidal particles is
6.53∙109 l mol
-1 s
-1.
1.7. The electrical double layer
In electrostatically stabilized suspensions, the surface charge of a colloidal particle
attracts ions of the opposite charge (counter-ions) that attach firmly, building the so-
called Stern layer. More counter-ions are then attracted in order to neutralize the charged
particle but these repel each other and are repelled by the ions in the Stern layer.
Therefore, they form a dynamic diffuse layer of counter-ions. The concentration of
counter-ions in the diffuse layer gradually decreases with distance from the surface, until
it reaches the concentration in the bulk. The Stern layer and diffuse layer constitute the
electrical double layer. The thickness of the electrical double layer depends on the type
and concentration of the ions in the suspension, the particle surface, temperature, etc. The
Debye-Hückel parameter κ is the inverse thickness of the electrical double layer defined
as:
2/1
0
0
22
Tk
nze
Br
iii
(9)
10
where e is the elementary charge, z is the ion charge, n0 is the number of ions per cubic
metre, εo is the permittivity of vacuum, εr is the dielectric constant, kB is the Boltzmann
constant and T is the absolute temperature.
1.8. Total inter-particle energy
The DLVO theory, named after Derjaguin, Landau, Verwey and Overbeek, describes the
interaction of two charged particles as the total energy (VT) that results from the sum of
the electro-osmotic repulsion between the ionic clouds and the Van der Waals attraction:
ART VVV (10)
The repulsive energy (VR) of the particles as a result of the interactions between their
diffuse-double layers can for spherical geometry be approximated by the expression:
kxB
R eTkan
V 22
2
064
(11)
where a is the particle radius, n0 is the number of ions per unit volume in the bulk, x is the
distance of interaction, κ is given by expression (9), and γ is a factor relating to the
surface potential (ψo) through the less restricted Gouy-Chapman expression (12), where z
is the valence of the electrolyte:
12
exp
12
exp
0
0
Tkze
Tkze
B
B
(12)
The attractive Van der Waals energy (VA) for spherical particles is given by the
expression:
2
2
2
2
2
2
2
4ln
2
2
4
2
6 ax
axs
ax
a
axx
aHVA (13)
where H is the Hamaker constant defined as:
2/32
2
2
1
22
2
2
1
2
21
21
216
3
4
3
nn
nnhTkH eB
(14)
where ε1 and ε2 are the dielectric constants for the material and the medium, h is Planck´s
constant, ve is the mean electronic adsorption frequency in the UV spectrum of the
11
medium and n1 and n2 are the refractive index in the visible spectrum for the material and
the medium, respectively.44,45
The total energy function VT displays a maximum at a certain particle distance. The
maximum of the total energy function corresponds to the energy that the particles must
surmount to aggregate. When the maximum of the total energy is negative, attraction
between particles will dominate and the system will be unstable, since every collision
between particles will lead to aggregation. On the contrary, when the total energy
maximum is positive, only the collisions with enough energy to overcome the energy
barrier will form an aggregate. Upon increasing the energy barrier, the fraction of
collisions with sufficient energy to overcome the energy barrier will be reduced, resulting
in a more stable suspension.
The height of the energy barrier depends on the surface potential and the electrolyte
concentration. Increasing electrolyte concentration reduces the double layer thickness, as
can be deduced from equation (9). The double layer compression reduces the repulsive
energy between particles according to equation (11). The electrolyte concentration at
which the repulsive energy is equal to the attractive energy is called the critical
coagulation concentration, CCC.46
At CCC and higher electrolyte concentrations,
colloidal suspensions are unstable. The height of the maximum total energy, which can
be compared with the activation energy for particle aggregation, is zero. Therefore every
collision between particles forms an aggregate, and the aggregation process takes place in
the diffusion-controlled regime.
1.9. Diffusive, advective (or conductive) and dispersive transport
The mechanisms for colloid transport through water-saturated geologic medium include
advection due to bulk motion of the fluid (also called convection) and dispersive
transport. Dispersive transport is caused by diffusion (similar to molecules) and
mechanical mixing by velocity variations in the porous matrix. Mechanical mixing
includes differences in fluid velocity in the center of the pores compared to the edges,
differences in pathway lengths and differences in velocity in larger and smaller pores.
Diffusion is the dominating transport mechanism at groundwater velocities lower than 1.6
10-10
m s-1
.47,48
The diffusion of colloidal particles in groundwater is described by the
Fick’s first law. The Fick´s first law establish the relationship between the mass flux per
unit time per unit area FDiff (kg m-2
s-1
) and the concentration gradient dC/dx, (kg m
-3)
where s is the distance in m and D is the effective diffusion coefficient of the medium
with units of m2·s
-1:
dx
dCDFDiff (15)
12
The negative sign in the above equation indicates that particle migration is in the
direction of decreasing concentration. The Fick’s first law is modified for a saturated
porous media according to the following expression:
dx
dCDF eDiff (16)
where φe is the effective porosity of the medium.
Total porosity (φ) is the ratio of volume of void space or interstices (Vv) to the total
volume (Vt).49
t
v
V
V (17)
Effective porosity (φe) is the volume that corresponds to interconnected pores through
which water can flow. Thus, not all the void space contributes to groundwater flow and
effective porosity is lower than total porosity since it does not include non-interconnected
pores.
Advective transport (or convective transport) occurs when colloidal particles move with
the water front. The flux for advection FAdv is expressed as:
CqCnvF xexAdv (18)
where FAdv represents the mass of colloid per unit cross-sectional area transported in the x
direction per unit time, qx is the Darcy velocity in the same direction,50
ne the effective
porosity, and vx the groundwater velocity or pore water velocity in the same direction.
Transported colloidal particles are subject to longitudinal or mechanical dispersion due to
velocity variations and differences in travel time along different flow paths. The
mechanical dispersive flux (or dispersive flux) FDisp can be described similarly to the
diffusive flux:
dx
dCDF emDisp (19)
where Dm is the mechanical dispersion coefficient.
The total flux FT in porous media will be the sum of the above introduced three flux FDiff,
FAdv and FDisp:
DispAdvDiffT FFFF (20)
Substituting equations (16), (18) and (19) on (20) we obtain:
13
dx
dCDCvF LexeT (21)
where DL is the longitudinal diffusion coefficient which corresponds to the sum of the
effective molecular diffusion coefficient D, and the Dm is the mechanical dispersion
coefficient:51
mL DDD (22)
The advective–dispersive (or convective-dispersive) transport is to some extent a
probabilistic process, since which flow line particles will step over at a grain boundary
follows statistics. Therefore, the concentration of particles breaking through after
transport in a porous medium describes a breakthrough curve rather than a narrow peak.52
1.10. Colloid retention processes in water bearing fractures
Particle sedimentation, attachment to surfaces and filtration are the main retention
mechanisms that prevent colloid transport. Figure 5 summarizes all the processes relevant
to colloid transport in natural fractures.
Figure 5: Colloid transport in natural fractures.
14
Suspended colloidal particles in groundwater flowing through a porous medium such as
fracture filling materials under advection regime might interact by sorption with the
mineral surfaces, become entrapped in pore constrictions and surface roughness
(filtration), disperse in the porous matrix (mechanical dispersion) or continue flowing
without capture.
Sorption is the general term used when the retention mechanism of ions or molecules
(sorbate) at a solid surface (sorbent) is unknown. Sorption includes adsorption, surface
precipitation and polymerization processes.53
The solid-water distribution coefficients (Kd) are widely used by geochemists to
determine the accumulation of solute (sorbate) at solid surfaces (sorbent) in an aqueous
medium. It should be noted that the determined Kd values do not provide information
about sorption mechanisms, that can be either by adsorption, surface precipitation or
polymerization, involving both physical and chemical processes.
For sorption processes independent of the concentration of sorbate, the distribution
coefficient, Kd, is a constant that expresses the ratio of equilibrium colloid concentrations
sorbed and in solution, according to the following expression:
m
V
C
CCK
f
fi
d
(23)
where Ci and Cf are the initial and final sorbate concentrations in solution, V is the
volume of solution in m3 and m is the mass of sorbent in kilograms.
By normalizing the Kd values to the specific surface area sa of the sorbent, comparison of
sorption between different sorbents becomes more straight forward.
af
fi
dsm
V
C
CCK
(24)
where sa is the surface area of the mineral in m2 kg
-1 usually determined by BET
measurements. Thus, the surface normalized Kd values have units of m.
15
2. Materials
2.1. Bentonite
Bentonites of different origin differ in exchangeable cations, montmorillonite content,
number and abundance of accessory minerals. The Wyoming form of bentonite (MX-80)
is widely used as reference material to investigate different properties of the buffer and
backfilling material in the deep repository. The bentonite used in this study is Wyoming
MX-80 supplied by the American Colloid Co.
The mineral composition of MX-80 is given in Table 1.
Table 1. Mineral composition of bentonite Wyoming MX-80.
18
Component MX-80
(wt-%)
Uncertainty
(± wt-%)
Calcite + Siderite 0–1 1
Quartz 3 0.5
Cristobalite 2 0.5
Pyrite 0.07 0.05
Mica 4 1
Gypsum 0.7 0.2
Albite 3 1
Dolomite 0 1
Montmorillonite 87 3
Na- 72% 5
Ca- 18% 5
Mg- 8% 5
K- 2% 1
Anorthoclase 0 1
CEC (meq 100 g-1
) 75 2
Organic carbon 0.2 –
Two different procedures were used to prepare colloidal bentonite suspensions.
For one series of experiments, (Paper III), suspensions were prepared from commercial
bentonite Wyoming (MX-80) as received from the supplier. One gram of bentonite was
dispersed in 250 ml of 10-3
M NaCl or NaClO4 solutions and the suspension was allowed
to settle for 24 h. Then 80 ml of the upper colloidal part were collected and used in the
experiments. The colloid concentration was determined gravimetrically to be 0.88 0.06
mg ml-1
. The pH of the suspensions was measured to approximately 9.
16
In other series of experiments (Paper IV), soluble and coarse minerals were removed
from bentonite as follows: (i) 0.3 g of clay was dispersed in 30 ml de-ionized water and
centrifuged for two hours at 6000 rpm. (ii) The supernatant solution was decanted off.
(iii) The sediment was collected and the clay fraction was separated from the coarser
minerals. (iv) The clay fraction was dispersed in 30 ml de-ionized water and centrifuged
at 6000 rpm for two hours. The conductance of the rejected supernatant was <80 μS. (v)
The sediment was collected and oven-dried at 60°C.
A stock colloid suspension of 0.8 g l-1
was prepared from the purified material. The initial
pH of the suspension was 8.3±0.1. Small volumes of NaOH was used to adjust the pH to
10 and 11. NaOH additions were buffered by bentonite and the pH of the suspensions
returned to the initial value. Daily additions of NaOH were necessary for at least two
weeks until the pH of the suspensions was stable. Once the basic pH was stable, the
acidic pH was adjusted to 3, 4 and 6 by adding small HCl volumes. These values were
achieved immediately after adding HCl and were constant so no further additions were
necessary. By conductance measurements, the concentration of NaOH was determined in
the suspensions with pH 10 and 11. The corresponding amount of NaCl was added to the
suspensions with pH 3, 4, 6 and 8.3 in order to obtain equal total ionic strengths. The
total ionic strengths studied were 0.002, 0.003, 0.005 and 0.0068 M. The pH of the
suspension with ionic strength 0.005 M was adjusted to 10 by using the buffer pair
NaHCO3/Na2CO3 instead of NaOH for comparison.
2.2. Sodium- and calcium-montmorillonite
The sodium- and calcium-montmorillonites used (Paper II, V, VI) were obtained by
purifying Wyoming bentonite MX-80 at the laboratory of Clay Technology AB. The
purification procedure was as follows: A 10 g portion of MX-80 bentonite was dispersed
in 1 L of 1 M analytical grade chloride solution of the desired cation and left to settle.
The supernatant was removed and the procedure repeated three times. The material was
washed three times with de-ionized water and the supernatant was removed after
centrifugation. The clay fraction suspension was separated from the accessory minerals
by decanting. In order to remove excess electrolytes, the clay suspension was transferred
to dialysis membranes (Spectrapore 3, 3500 MWCO) placed in plastic containers with
de-ionized water. The water was changed daily until the electrical conductivity was
below 10 μS cm-1
. The material was then redispersed in 1 L of 1 M analytical grade
chloride, and the process was repeated again. The montmorillonite was then dried at 60oC
and milled to an aggregate grain size similar to that of MX-80. Thereafter, the structural
formula of the Na-exchanged montmorillonite was determined by ICP-AES element
analysis as:18
(Na 0.64 K 0.01)(Al 3.11 Ti 0.01 Fe 3+
0.36 Mg 0.47)[( Si 7.93 Al 0.07)O20](OH)4 x nH2O (25)
and that of the Ca-exchanged form as:
17
(Ca 0.25 Na 0.01 K 0.01)( Al 3.14 Ti 0.01 Fe 3+
0.37 Mg 0.47 )[(Si 7.93 Al 0.04) O20](OH)4 x nH2O
(26)
Purified sodium- and calcium-montmorillonite were used to prepare suspensions of 0.8 g
L-1
.The dry material was poured into deionised water and the suspension was put into an
ultrasonic bath for two hours (Paper II). The pH measured in the suspensions was 8.3 ±
0.1 The ionic strength was adjusted by addition of small volumes of concentrated NaCl
and CaCl2 solutions.
In sorption experiments (Paper V), 30 mL of Na-montmorillonite colloidal suspensions
with a concentration of 5 x 10-4
g L-1
were in contact with different amounts of quartz
sand using buffered solutions 0.001 M of TRIS (tris(hydroxymethyl)aminomethane) at
pH 8.5 and MES (2-(N-morpholino)ethane-sulfonic acid) at pH 6 and 4.5.
For transport experiments performed in Paper VI, 0.05 g of dry clay was treated
ultrasonically until complete dispersion in 2 mL deionized water containing the
conservative (non-sorbing) tracer tritium (HTO) 600 Bq cm-3
(TRY64, 2 MBq, 5 g,
Canberra).
2.3. Reference colloids
Carboxyl and amidine terminated surfactant-free polystyrene latex particles purchased
from Interfacial Dynamics Corporation (IDC, Portland, USA) were used as reference
colloids in stability tests at different temperatures (Paper III). Suspensions were prepared
by diluting the stock supplied to particle concentrations in the range 0.38-3.10 mg mL-1
in
10-3
M NaCl or NaClO4 solutions.
For sorption and transport experiments (Papers V VI), fluorescent polystyrene latex
colloidal particles (Duke Scientific Corporation, CA, USA) were used. Commercial
names G50B, R100B and B200B correspond to 50±5, 100±5 and 200±10 nm diameter
respectively. Batch sorption experiments were performed in 0.001 and 0.01 M buffer
solutions, using TRIS at pH 8.5 and MES at pH 6 and 4.5.
Table 2 provides size, zeta potentials of reference and natural colloids.
The concentration of fluorescent particles in batch sorption and in colloid transport
experiments was (5±0.2) x 109 particles mL
-1 for 50 nm, (1.0±0.2) x 10
10 particles mL
-1
for 100 and 200 nm respectively.
The latex colloidal cocktail was introduced in the column as a 1 mL single injection
containing and tritium (HTO) 600 Bq cm-3
used as conservative tracer.
18
Table 2. Diameter and zeta potential of the colloidal particles at 0.001 M.
Diameter
(nm)
Temperature
(ºC)
Zeta potential (mV)
pH=8-8.5 pH=6 pH=4.5
Bentonite 20-600 0-20-80 -508a
Na-mont.+NaCl 20-600 20 -54±9a
Ca-mont.+NaCl 20-600 20 -47±8a
Na-mont.+CaCl2 20-600 20 -25±5a
Ca-mont.+CaCl2 20-600 20 -19±4a
Amidine Latex 490±11 20 +7410a
Amidine Latex 490±11 80 +74a
Fluorescent Latex 50±5 22 -33±1b -31±1
b -22±1
b
Fluorescent Latex 100±5 22 -45±2b -45±2
b -45±2
b
Fluorescent Latex 200±10 22 -50±2b -50±2
b -42±2
b
Na-montmorillonite 20-600 22 -38±2b -41±2
b -41±2
b
a Measured by means of a ZetaPALS Zeta Potential Analyzer, Brookhaven Instruments Corporation
b Measured by means of a Malvern Zetasizer 3, Malvern Instruments.
2.4. Minerals
The minerals used in Paper V and VI were quartz, granite and the fracture filling
materials epidote, biotite and calcite.
Quartz sand was purchased from Fluka, 99.999% SiO2. Quartz with BCR reference
material No.131, and a size range of 0.48-1.8 mm, was used in sorption experiments
(Paper V) and quartz BCR reference materials No.068 and No. 132, with size ranges of
0.16-0.63 and 1.4-5 mm respectively was used in transport experiments (Paper VI).
Granite, epidote and calcite were sampled at the Äspö Hardrock Laboratory in Sweden
and grinded. The 0.5-2 mm fraction was used in sorption experiments.
Characterization of granite, epidote and calcite obtained at Äspö Hardrock Laboratory
was carried out by powder X-ray diffraction, XRD (X´Pert PRO, PANalytical B.V., The
Netherlands). The composition of the different mineral samples is presented in Table 3.
As can be seen in Table 3, some minerals consist of a pure single phase like calcite, while
others have different minerals phases.
The biotite used in sorption and transport experiments (Paper V, VI) was sampled in
Helle, Norway and kindly provided by The Royal Museum of Natural History,
Stockholm, Sweden; (catalog number LK-5210).
19
The biotite used in the experiments was characterized elsewhere.54
Powder X-ray
diffraction analysis revealed that the specimen consisted of biotite without detectable
impurities.
The biotite specimen was cleaved manually with pincers in order to obtain fresh new
biotite particles. The particle size ranged from 1 to 3 mm in diameter. For transport
experiments (Paper VI), different amounts of biotite (10-20g) were used to pack the
columns mixed with finest quartz sand.
All the minerals were repeatedly washed with milli-Q water to remove colloidal particles
prior to the experiments.
Table 3. Structural formula, pzc and phase occurrence (%) of the minerals used in this study.
Mineral % Formula pzc
Quartz Quartz 100 SiO2 2 53
Granite Quartz 39 SiO2 2 53
Albite 31 NaAlSi3O8 2 53
Biotite 29 K(Mg,Fe)3 AlSi3O10 (OH,F)2 6.5 55
Epidote Epidote 30.7 Ca2(Fe3+
Al)3 (SiO4)3 (OH)
K-Feldspar 38 KAlSi3O8 2-2.4 53
Chlorite 21 (Mg,Fe,Al)12 (Si,Al)8 O20
(OH)16 5
56
Quartz 10.2 SiO2 2 53
Biotite 100 K(Mg,Fe)3 AlSi3O10 (OH,F)2 6.5 55
Calcite 100 CaCO3 ---
The morphology and surface roughness of the different minerals was examined by using
Scanning Electron Microscopy (SEM), using a JEOL JSM9460LV SEM/EDS
microscope, Japan. Pictures of the surface roughness of fine and coarse quartz are shown
in Figure 6.
20
Figure 6: Scanning electron microscopy image of the fine quartz sand (0.16-0.63 mm) and the
coarse quartz sand 1.4-5 mm (non-coated sample, 10 kV accelerating voltage). The scale bars is 5
µm.
The specific surface areas of the different minerals were determined by the Brunauer-
Emmett-Teller (BET) isotherm using N2 as adsorbing gas (Flowsorb II 2300,
Micromeritics, USA). The resulting specific surface areas are listed in Table 4.
Table 4. Size fraction and specific surface area of minerals.
Mineral Size fraction (mm) Specific Surface Area (m2 kg
-1)
Quartz fine 0.16-0.63 190±50
Quartz medium 0.48-1.8 340±50
Quartz coarse 1.4-5 560±50
Granite 0.5-2 410±50
Epidote 0.5-2 620±50
Calcite 0.5-2 320±50
Biotite 1-3 130±50
21
3. Methods
3.1. Single Particle Counting
Single Particle Counting (SPC) is used to count and size the particles present in water by
analyzing the particle-scattering of the laser light for a range of angles. Pulse height
analysis is used for sizing the scattered light into a number of size channels. The
instrument consists of a pump, a syringe, a laser-detector unit and a computer for data
processing.57
The HSLIS-M50 unit of the SPC measures concentration of colloidal particles in four
size channels: 50-100, 100-150, 150-200 and >200 nm. HVLIS-C200 unit measures
colloid concentrations in the larger size region, consisting of eight channels: 200-300,
300-500, 500-700, 700-1000, 1000-1500, 1500-2000, 2000-5000 and >5000 nm.58
The pump ensures a constant milli-Q water flow. The syringe is used to inject the
samples into the on-line flow and the light-scatter phenomenon is analyzed.
Given the high sensitivity of the instrument and to ensure that only one scattering event
takes place when one particle passes the laser beam, sample dilution is often needed.
3.2. Photon Correlation Spectroscopy
Photon correlation spectroscopy (PCS) measures the size of particles in solution in the
size range 10-2000 nm by determining the diffusion coefficient. The advantages of this
technique are that it is fast, non-invasive and requires only a small sample volume.
Photon correlation spectroscopy measures the dynamic light scattering (or quasi-elastic
light scattering). When a particle is illuminated by the laser beam, the phase of the
scattered light is dependent on its position. The total intensity of the scattered light is the
result of all the individual scattered waves. Fluctuations arise in the scattering intensity at
a given scattering angle because the phase and polarization of the light scattered by each
particle alter over time as the position of the particle changes due to the Brownian
motion. The fluctuations over time of the scattered light have a lifetime τ that is recorded
and analyzed using an autocorrelation function. The average of all the pulses in small
time intervals gives the intensity of the autocorrelation function C(τ):
BKDAC T )exp()( 2 (27)
where A and B are constants, DT is the translational diffusion coefficient, and K is a
constant calculated as:
)2/sin()/4( 0 nK (28)
22
where n is the refractive index of the liquid, λo is the wavelength of the laser beam and θ
is the scattering angle. Since the diffusion rate of particles is determined by their size, the
rate of fluctuation of the scattered light can be used to determine their size. The Stokes-
Einstein equation gives the relationship between the hydrodynamic diameter d and the
diffusion coefficient DT:
dTkD bT 3/ (29)
where kB is the Boltzmann constant and T is the temperature.
The signal of the PCS instrument, given as counts per second, is mainly determined by
the number of particles scattered. However, the refractive index and the geometry of the
particles can also affect the intensity of the signal.59
In this study, the PCS instrument used was a 90Plus Particle Sizer supplied by
Brookhaven Instruments Corporation. The instrument consists of a He-Ne laser source, a
set of optical elements to collimate, focus and polarize the beam, a sample cell placed in a
temperature-controlled module, a second set of optical elements to collect the scattered
light, an amplifier and a detection system that counts the number of photons occurring in
a defined time interval, a correlator that stores the counts and fits the time average
calculations to a correlation function and a computer for parameter input and data
output.60
In order to determine colloid aggregation rates, it is necessary to quantify the number of
particles in suspension. One possibility is to count each particle in suspension, for
example by using SPC, while another alternative is to determine one property of the
colloidal system that is proportional to the concentration of paticles in suspension, which
is the case when using the PCS technique.
The signal given by PCS can be used as a relative measure of the concentration of
particles under certain conditions. If the size distribution and the mean size of the
particles do not significantly change over time, the change in the count rate over time is
only dependent on the change in the concentration of particles. In a colloidal suspension,
the concentration of particles decreases with time due to aggregation and subsequent
sedimentation of large aggregates. At sufficiently low electrolyte concentrations and/or
low particle concentration, the aggregation rate is slower than the sedimentation rate.
When aggregation is the rate-determining step, the larger particles move downwards
while the smaller ones remain in the upper part of the suspension. The mean size and the
size distribution of the particles in the upper part of the suspension do not change
significantly over time. Therefore, the PCS signal, given as the count rate of the
measurement, can be used as a relative measure of the particle concentration.
23
3.3. Zeta potential analysis
The zeta potential is the value of the surface potential ψ at the Stern layer plane where the
diffuse electrical double layer of the particle starts. It can be determined from the
movement of the charged particles in the presence of an electrical field. Depending on the
sign of the charge, the particles move either to the positive or to the negative pole. The
velocity of the movement is proportional to the charge of the particle. The electrophoretic
mobility, ue, of the particles can be expressed as:
)(3
2af
E
vue
(30)
where v is the particle velocity, E is the electric field, ζ is the zeta potential in mV, ε is the
dielectric constant of the medium, η is the viscosity of the medium, f(κa) is the Henry
function, being κ the inverse thickness of the electrical double layer and a the radius of
the particles.61
The function f(κa) varies from 1 to 1.5 depending on ka. For large particles with a thin
double layer, where f(κa) is >> 1 and f(κa)=1.5, the electric field does not affect the
mobility of the particles and equation (31) is known as the Smoluchowski equation. In
the case of small particles in diluted aqueous solution, κa <1 and f(κa)=1, the ions in the
double layer surrounding the particle also move due to the electric field but in the
opposite direction to the particle, which causes a reduction in particle velocity.
Expression (31) is then known as the Hückel equation.62
A ZetaPALS Zeta Potential Analyzer supplied by Brookhaven Instruments Corporation
or a Malvern Zetasizer 3 from Malvern instruments were used to determine the zeta
potential of the particles. The acronym PALS stands for Phase Analysis Light Scattering.
Two electrodes provide an electrical field. The light of laser beam is scattered by the
particles. Since the particles are in movement, the Doppler effect of the scattered light is
used to calculate the velocity of the particles.
3.4. Fluorescence spectrophotometry
Fluorescent substances are able to absorb a photon of a given wavelength which induces
an electronic energy transition to a higher state without electron spin change. When the
excited electrons return to the ground state, the emitted radiation has usually a longer
wavelength than the radiation absorbed for excitation. The shift during emission to longer
wavelengths is known as the Stokes shift.
Qualitative and quantitative information of fluorescence substances can be obtained by
analyzing the emission spectra after excitation, considering the linear relationship
between the power of fluorescent radiation, F, and the concentration of the fluorescence
substance, c, as given by the following expression:
24
bcPKF 303.2´ 0 (31)
where P0 is the power of the incident beam, is the molar absorptivity of the
fluorescence substance, K´ is a constant depending on the quantum efficiency of the
process, and b is the length of the medium.63
The concentration of fluorescent polystyrene latex colloids was determined by means of a
Cary Eclipse Fluorescence Spectrophotometer, Varian, CA, USA. The instrument is
provided with a full spectrum Xe pulse lamp, a monochromator for excitation wavelength
selection, a second monocromator, a photoelectric detector and a computer for
operation.64
3.5. Liquid scintillation counting
In a beta decay process, the nucleus emits an electron (β- or e0
1 ), a positron (β+ or e0
1 ) or
undergoes electron capture. These three processes can be represented as:
0
11XX A
Z
A
Z (32)
0
1
0
11
0
11 eXXX A
Z
A
Z
A
Z (33)
XX A
Z
A
Z 1 (34)
where X represents the atom, A is the mass number or number of nucleons (protons and
neutrons), Z is the atomic number, v symbolizes the neutrino and represents the
antineutrino.
For low energy beta emitters (C-14, S-35, H-3) liquid scintillation is the most
appropriated detection technique. The sample containing the beta emitter is dissolved in
the scintillating solution (Ready Safe, Beckman) and placed between two photomultiplier
tubes of a beta counter Tri-Carb 1500, Packard. The energy of the beta particle is
absorbed by the scintillation liquid and light is emitted. The signal is received by the
photomultipliers and analyzed by a coincidence counter that only records when two
pulses from the photomultipliers arrive simultaneously.65
25
4. Results and Discussion
4.1. Colloid generation
It is known that the amount of colloidal particles generated increases with increasing flow
and that colloidal particles are generated at flow rates as low as 0.49 mL d-1
and under
quasi-static conditions.66,67
When a pellet of compacted bentonite is in contact with water,
the attractive and repulsive forces in the swollen montmorillonite front are in equilibrium.
The higher the swelling pressure, the more colloidal particles are generated due to higher
repulsion between aluminosilicate layers as was shown by Missana et al. 200366
If
colloidal particles are removed by flow shear forces or by self diffusion, new particles are
generated to re-establish the equilibrium.
In this thesis colloid generation in the absence of flow was studied. Under static
conditions, colloidal particles are not eroded from the source by shear force. The only
mechanism for colloid formation is the swelling of the clay, followed by particle
detachment and subsequent self-diffusion of colloidal particles from the colloid source
toward the solution.
The colloid concentration in equilibrium is the result of a cyclic process consisting of
particle detachment from the source, diffusion, aggregation, sedimentation and again
detachment. Thus, one could describe the population of colloidal particles in suspension
as dynamic since new particles are released and settle continiously. A sketch
summarizing all these processes is represented in Figure 7.
Figure 7: Colloidal life cycle.
26
Figure 8 shows photographs of generation-sedimentation experimental set-up. In the
generation experiments, the dry compacted pellet is in contact with water, and particle
detachment and diffusion take place in preference to aggregation and sedimentation. On
the other hand, in a newly-prepared colloid dispersion, aggregation and sedimentation
take place faster and the concentration of particles in suspension decreases over time.
Figure 8: Generation (a, b) and sedimentation (c, d), before (a, c) and after (b, d) the
experiments.
d c
b a
27
For both types of experiments, the final concentrations of particles in suspension are
found to be very similar, as shown in Figure 8b and 8d. This can be understood by taking
into consideration that all these processes are controlled by electrostatic forces. Thus,
when the system reaches an apparent local steady state concentration, the rate of
generation is identical to the rate of aggregation. Since the concentration of particles is
not homogeneous along the test tubes, this has been referred to as “pseudo-
equilibrium”.67
In order to quantify mass transport from the bentonite barrier under different conditions,
it is necessary to know the equilibrium concentrations of colloids outside the barrier. The
results from simple sedimentation experiments are compared to those from generation
experiments.
50-100 nm
100-150 nm
150-200 nm
200-300 nm
300-500 nm
500-700 nm
700-1000 nm
1000-1500 nm
1500-2000 nm
2000-5000 nm
Figure 9: Normalised Na-montmorillonite colloid concentrations for generation (a) and
sedimentation (b) experiments as a function of time at NaCl concentration 0.001 M. Samples
taken in the middle of the batch.
With increasing ionic strength, the equilibrium particle concentration decreases and
pseudo-equilibrium is reached more rapidly. There are two reasons for this behavior:
first, the swelling pressure decreases with increasing ionic strength,40
which reduces
colloid generation and second, as the ionic strength increases, the double layer
compresses according to Eq. (9) and aggregation becomes faster due to less repulsion
between particles. The influence of ionic strength is discussed in more detail in section
4.2.
NaCl 0.001 M
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
1.E+08
1.E+09
0 1 10 100 1000 10000Time (h)
Co
llo
id c
on
cen
tra
tio
n (
ml
-1 n
m-1)
a
Generation
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
1.E+08
1.E+09
0110100100010000Time (h)
Co
llo
id c
on
ce
ntr
atio
n (
ml
-1 n
m-1)
NaCl 0.001 M
b
Sedimentation
28
Figure 10: Pseudo-equilibrium profiles in generation and sedimentation experiments at time zero
and after 7000, 6000 and 2300 hours for Na-montmorillonite in 0.001, 0.01 and 0.1 M NaCl
respectively and after 1500 hours for Ca-montmorillonite in 0.001 M NaCl.
The concentration of colloids in mg L-1
in the size range 50-5000 nm can be estimated
from the pseudo-equilibrium profiles shown in Figure 10.
Assuming spherical geometry for colloidal montmorillonite particles with a density of 1.5
g cm-3
, the mass in grams of colloids at pseudo-equilibrium can be approximated by
multiplying the density of the particles by the number of colloidal particles and the
average volume in each size fraction as expressed in equation (35):
11 i
particleiiparticlei
i
iTotal VNmNm (35)
where mTotal is the concentration of montmorillonite in the suspension expressed in g L-1
,
Ni is the number of particles per litre at pseudo-equilibrium for each size fraction as
determined by SPC, mi particle is the average mass of the colloidal particle for a given size
fraction, ρ is the particle density and Vi particle is the average volume of the particles in a
given size fraction. The radius of the particles was determined as the average of each size
fraction. Due to these approximations, the mass of colloids is a rough estimate rather than
an exact value.
In the case of Na-montmorillonite in contact with 0.001 M NaCl solution, the
concentration of particles in suspension was determined as an average of the colloid
concentrations in generation and sedimentation experiments, since this system has not
reached pseudo-equilibrium.
29
Table 5 summarises the pseudo-equilibrium concentrations of colloidal Na- and Ca-
montmorillonite for each ionic strength studied.
Table 5. Pseudo-equilibrium colloid concentrations (mg L
-1) for Na- and Ca-montmorillonite in
the size range 50-5000 nm in contact with different concentrations of NaCl solution.
NaCl (M) 0.001 0.01 0.1
Na-montmorillonite 5.2±0.5 0.5±0.1 0.2±0.1
Ca-montmorillonite 0.4±0.2 --- ---
From Table 5 we can see that colloids are present in suspension despite the relatively
high salinity in the medium. Regarding the first question in the sketch presented in Figure
2, it becomes clear that colloid generation is a relevant process that should not be
disregarded. Colloid generation may gain even greater importance in a flowing system,
where physical erosion of colloidal particles will increase the colloid population.
Even if generation experiments with Ca-montmorillonite were performed in NaCl
solutions and cation exchange reactions took place, the colloid pseudo-equilibrium
concentrations differ by one order of magnitude between Na- and Ca-montmorillonite.
This indicates that cation exchange is only partial.
4.2. Stability of colloidal suspensions
A colloidal suspension is said to be stable if the particles remain in suspension during
over a long period of time, showing low tendency to aggregate. The concept of stability is
arbitrary since it depends on time as such and the time frame of other processes in the
system.
4.2.1. Method applicability and PCS calibration
Figure 11 shows the change in the mean size of different cationic forms of
montmorillonite particles dispersed in NaCl and CaCl2 electrolyte as a function of time.
As can be seen in Figure 11, variations in the relative mean size over time are not
significant. The reason for this is that large aggregates settle and are not detected by PCS
in this experimental set-up.
30
0
50
100
150
200
250
300
350
400
450
0 10000 20000 30000 40000 50000 60000
Time (min)
Siz
e (
nm
) .
NaMont+NaCl
CaMont+CaCl2
Figure 11: Mean size of Na- and Ca-montmorillonite particles as a function of time.
The PCS signal is sensitive to temperature. In Figure 12, the signal given by the
instrument against the particle concentration is represented for three different
temperatures.
0
0.2
0.4
0.6
0.8
1
1.2
0 0.2 0.4 0.6 0.8 1 1.2
Concentration mg ml-1
Co
un
t R
ate
(M
cp
s
-1 )
.
Figure 12: Relationship between the PCS count rate and the concentration of bentonite particles
at pH 9 and 0C (), 20C (), and 80C ().
The count rates were linearly proportional to the concentration of particles for the three
temperatures (Figure 12). Hence the calibration curves are used to quantitatively
determine colloid concentrations. However, the signal obtained at 80ºC was lower than at
0 and 20ºC. Therefore, the readings at the highest temperature were normalized by
multiplying by the factor of proportionality between 20 and 80ºC.
31
4.2.2. Effect of the ionic strength on aggregation kinetics
One of the groundwater parameters that most strongly determine the stability of a
colloidal suspension is ionic strength. This is easy to understand bearing in mind the
dependency of the repulsive force on ionic strength according to equations (9)-(14).
To evaluate the impact of ionic strength on colloid aggregation, the kinetics of
aggregation was studied by using the data treatment described in section 4.2.1. This
method allows the determination of the rate constants for aggregation at different ionic
strengths.
Plotting the inverse of the count rate versus time gave straight lines with correlation
factors R2
≥ 0.9, which indicates that the aggregation of the colloidal particles studied
(bentonite, montmorillonite and amidine latex) follows second order kinetics, as
expected.42
The aggregation kinetics of Na- and Ca-montmorillonite dispersed in solutions with
different concentrations of NaCl or CaCl2 were studied. For example Figure 13 shows the
aggregation kinetics of montmorillonite particles at different NaCl concentrations.
Na-Montmorillonite in NaCl
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 10000 20000 30000 40000 50000 60000
Time (min)
1/C
R (
Mcp
s
-1
) .
3*10-3 M
4*10-3 M
5*10-3 M
6*10-3 M
7*10-3 M
8*10-3 M
9*10-3 M
10-2 M
Figure 13: Inverse of the count rate versus time for Na-montmorillonite at different
concentrations of NaCl.
In Figure 13 it can be observed that increasing electrolyte concentration accelerated
aggregation kinetics. Similar results were obtained using CaCl2 as electrolyte.
The true second order rate constants were estimated from the second order slopes,
expressed in Mcps-1
min-1
units, by transforming the signal of the PCS instrument, given
as count rate, to particle concentration. The number of particles in a suspension can be
expressed as:
particleTotal mNm (36)
32
where mTotal is the mass of montmorillonite in the suspension expressed in g L-1
, mParticle
is the mass of each colloidal particle and N is the number of particles per litre. For the dry
mass of montmorillonite present in the experiments mTotal the count rate given by PCS
can be related to the number of particles N and the mParticle which was deduced from the
density once the volume of the particles was known. The volume of the particles was
estimated from the particle size data obtained from PCS assuming spherical geometry and
monodisperse suspensions (Paper II). The average diameter of Na-montmorillonite
particles was found to be 220 nm, while that of Ca-montmorillonite particles was 380 nm
(Figure 11). A density of 1500-1600 kg m-3
for hydrated particles of Na-montmorillonite
and 1700-1800 kg m-3
for Ca-montmorillonite reported by Pusch (2001)68
were used in
the calculations. The experimentally determined second order slopes and the
corresponding estimated second order rate constants are summarized in Table 6 for NaCl
electrolyte and in Table 7 for CaCl2 electrolyte.
Table 6. Second order slopes and second order rate constants for the aggregation kinetics of Na-
and Ca-montmorillonite particles at different concentrations of NaCl.
NaCl (M)
Na-montmorillonite Ca-montmorillonite
Slope (Mcps-1
min-1
) × 105
k (l mol-1
s-1
)
× 10-5
Slope (Mcps-1
min-1
) × 105
k (l mol-1
s-1
) ×
10-5
3×10-3
3.2 ± 0.4 0.09 ± 0.01 13 ± 1 4.6 ± 0.4
4×10-3
4.5 ± 0.3 0.118 ± 0.008 16 ± 2 5.7 ± 0.9
5×10-3
4.9 ± 0.2 0.134 ± 0.006 19 ± 2 6.0 ± 0.6
6×10-3
5.4 ± 0.2 0.147 ± 0.006 29 ± 1 8.2 ± 0.3
7×10-3
6.4 ± 0.4 0.172 ± 0.05 40 ± 5 11 ± 1
8×10-3
7.1 ± 0.2 0.194 ± 0.007 50 ± 10 14 ± 3
9×10-3
15.8 ± 0.7 0.43± 0.02 80 ± 20 19 ± 5
1×10-2
23.3 ± 0.9 0.63 ± 0.02 90 ± 10 23 ± 3
Table 7. Second order slopes and second order rate constants for the aggregation kinetics of Na-
or Ca-montmorillonite particles at different concentrations of CaCl2.
CaCl2 (M)
Na-montmorillonite Ca-Montmorillonite
Slope (Mcps-1
min-1
) × 105
k (l mol-1
s-1
) ×
10-5
Slope (Mcps-1
min-1
) × 105
k (l mol-1
s-1
) ×
10-5
3×10-4
4.5± 0.2 0.163 ± 0.009 22.1 ± 0.8 11.2 ± 0.4
4×10-4
7.3± 0.4 0.31 ± 0.02 47 ± 2 24 ± 1
5×10-4
26 ± 3 0.9 ± 0.1 54 ± 8 27 ± 4
6×10-4
37 ± 3 1.3 ± 0.1 103 ± 15 52 ± 8
7×10-4
155 ± 48 5 ± 2 166 ± 29 80 ± 14
8×10-4
218 ± 39 8 ± 1 197 ± 39 100 ± 20
9×10-4
400 ± 132 14 ± 5 276 ± 23 140 ± 9
As can be seen in Tables 6 and 7, the rate constants for aggregation increase with
increasing electrolyte concentration for both cationic forms of montmorillonite and
electrolytes, in agreement with DLVO theory.
33
When the data listed in Table 6 and Table 7 was plotted as the logarithm of the second
order rate constants versus the square root of the ionic strength of the suspensions, a
linear relationship was obtained (Figure 14).
Figure 14: Effect of electrolyte concentration on the aggregation kinetics of Na- and Ca-
montmorillonite for a) NaCl and b) CaCl2. CE indicates the electrolyte concentration necessary to
replace 98% of the exchangeable cations in montmorillonite.
In Figure 14, the ionic strength denoted by CE indicates the electrolyte concentration
required to exchange 98% of the calcium in Ca-montmorillonite dispersed in NaCl for
sodium (Figure 14a), and the concentration of CaCl2 necessary to replace 98% of the
sodium in Na-montmorillonite for calcium (Figure 14b). The CE quantities were
calculated from the cation exchange capacity of montmorillonite18
using the equilibrium
constants reported by Tang et al. 199369
for 98% replacement of the exchangeable
cations. Since cation exchange is an equilibrium reaction, a larger degree of replacement
occurs in the interlayer space as the concentration of electrolyte in the medium increases.
In Figure 14a, the calculated CE is lower than the concentrations used in the experiments.
Therefore, replacement of calcium by sodium in Ca-montmorillonite is expected to be
y = 91.821x + 1.379 R 2 = 0.984
y = 49.292x + 4.601 R 2 = 0.983
0
1 2
3
4
5
6 7
8
0 0.01 0.02 0.03 0.04 0.05 0.06
I 1/2
Lo
g k
Na-Montmorillonite
Ca-montmorillonite
CE= 0.26 b
y = 16.765x + 2.9556
R 2 = 0.8351
y = 16.043x + 4.7206 R2 = 0.9653
0
1
2
3
4
5
6
7
0 0.02 0.04 0.06 0.08 0.1 0.12
I 1/2
Lo
g k
Na-Montmorillonite
Ca-Montmorillonite
CE a
34
complete for the range of NaCl concentrations used in the experiments. Since both
materials are Na-montmorillonite, they show the same dependency on the logarithm of
the rate constant with the square root of the ionic strength. However, in the experiments
performed with CaCl2 (Figure 14b), Na- and Ca-montmorillonite showed different rates
of increase in log (k) with increasing ionic strength. The calculated CE value is much
higher than the CaCl2 concentrations used in the experiments. Thus, cation exchange
reactions can be expected to take place in the range of ionic strengths investigated. A
larger replacement of sodium by calcium in Na-montmorillonite occurred as the
concentration of CaCl2 increased.
The zeta potential of montmorillonite at buffer pH and room temperature was
independent of electrolyte concentration, but strongly dependent on the cationic form of
montmorillonite and the cation of the electrolyte. The average zeta potential values of
different electrolyte concentrations are given in Table 8.
Table 8. Average values of zeta potential of Na- and Ca-montmorillonite in de-ionized water and
NaCl or CaCl2 electrolyte.
Deionised water
(mV) NaCl (mV) CaCl2 (mV)
Na-montmorillonite -55 ± 2 -54 ± 9 -25 ± 5
Ca-montmorillonite -18 ± 2 -47 ± 8 -19 ± 4
The valence of the cation adsorbed in the Stern layer strongly determines the zeta
potential of the particles. Consequently, the charge of the montmorillonite colloid is
completely governed by the electrolyte in the medium.
DLVO calculations were performed in order to interpret the experimental results from
aggregation kinetics studies. The differences in zeta potential for the pure sodium and
pure calcium systems were taken into account when calculating the repulsive, attractive
and total energy using equations (9) to (14), assuming the zeta potential values to be
proportional to the surface potential of the particles.
The maximum values reached by the total energy are plotted against the square root of
the ionic strength in Figure 15.
35
y NaCl = -6.494E-18x + 4.470E-18
R2 =0.998
yCaCl2 = -8.513E-18x + 2.616E-18
R2 = 0.998
0
5E-19
1E-18
1.5E-18
2E-18
2.5E-18
3E-18
3.5E-18
4E-18
4.5E-18
0 0.05 0.1 0.15 0.2
I ½
VT
ma
x (
J)
.
NaCl
CaCl2
Figure 15: Maximum values of the total energy versus square root of the ionic strength. The solid
lines show linear extrapolations of the values calculated at the low ionic strengths inside the
square.
As can be seen in Figure 15, the height of the total energy function decreased with
increasing electrolyte concentration, which is in agreement with the kinetic results. For a
limited ionic strength range, such as the ionic strength range investigated experimentally
marked by a square in Figure 15, the maximum of the function decreases linearly with the
square root of the ionic strength. However, it deviates from linearity for a wider ionic
strength range. This explains the linear correlation between the logarithm of the rate
constant and the square root of the ionic strength observed for the fairly narrow ionic
strength range investigated experimentally (Figure 14a and 14b). Note that the total
energy maximum is more sensitive to increasing CaCl2 than increasing NaCl ionic
strength, which is in agreement with experimental results shown in Figure 14.
The linear relationships shown in Figure 14a and 14b allowed CCC to be estimated by
extrapolation to diffusion-controlled conditions. The diffusion-controlled rate constant
for particle aggregation calculated from equations (6)-(8), 6.53 x 109 l mol
-1 s
-1, was
inserted into the linear regression for the log (k) with I obtained experimentally for the
pure sodium and calcium systems. The ionic strengths obtained in this way correspond to
apparent CCC values. However, since log (k) is expected to be linearly related to VT max
and VT max is not strictly linearly related to I , the apparent CCC values must be
corrected. The correction is performed by identifying the I value at which the VT max
value corresponds to the diffusion-controlled limit using DLVO theory. The correction
procedure is illustrated in Figure 16 for the case of CaCl2.
36
y = -8.513E-18x + 2.616E-18
R2 = 0.998
0
5E-19
1E-18
1.5E-18
2E-18
2.5E-18
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
I ½
VT
ma
x (
J)
. .
CaCl2
CCC
corrected
CCC
apparent
Figure 16: Correction of the CCC value for Ca-montmorillonite in CaCl2 electrolyte.
The corrected CCC for Na-montmorillonite in NaCl was found to be 0.70 M and the
corrected CCC for Ca-montmorillonite in CaCl2 solutions was 0.0062 M. These CCC
values are of the same order of magnitude or higher than the CCC values published in the
literature, as can be seen in Table 9.
Table 9. Published CCC values for montmorillonite compared with those obtained in the present
study (Paper II).
Author (year) Mineral Clay/water
(%) Method CCC (M)
Frey (1978)71
Wyoming 0.025 Test tube series
(24h) 0.01-0.15 NaCl
Tombácz (1989)71
Na-Mont 0.2 Visual inspection
(24h)
0.25 NaCl
0.002 CaCl2
Tombácz(1989) 71
Ca-Mont 0.2 Visual inspection
(24h)
0.25 NaCl
0.002 CaCl2
Chheda (1992) 72
Na-Mont 0.015 Turbidity and
optical density
≈ 0.018 NaCl
≈0.00045 CaCl2
Lagaly (2002) 73
Na-Mont 0.025 Visual inspection 0.005 NaCl
0.0004 CaCl2
Lagaly (2002) 73
Na-Mont 0.5 Rheological
measurements
0.015 NaCl
0.002 CaCl2
Lagaly (2002)73
Na-Mont 1 Rheological
measurements
0.020 NaCl
0.003 CaCl2
SKB (2004a) 74
Mont Not
specified Not specified
0.1 NaCl
0.001 CaCl2
Paper II Na-Mont 0.08 Aggregation
kinetics 0.70±0.05 NaCl
Paper II Ca-Mont 0.08 Aggregation
kinetics
0.0062±0.0005
CaCl2
37
The CCC values determined by aggregation kinetics can be expected to be higher, since
other experimental methods (listed in Table 9) identify the CCC value with changes in
the colloidal system that become significant at aggregation rate constants far below the
point where the system begins to be diffusion-controlled. It is interesting to note that
under diffusion-controlled conditions, the total energy is still positive, when according to
the CCC definition given by DLVO theory the total energy should be zero. Therefore,
according to the DLVO theory, the CCC value should correspond to an even higher
electrolyte concentration.
The difference between the CCC values for NaCl and CaCl2 of two orders of magnitude
is mostly due to the differences in zeta potential, size and density between the Na- and
Ca-montmorillonite particles. However it would be of interest to use this method to
determine CCC values for colloidal systems not affected by cation exchange processes.
4.2.3. Influence of temperature on aggregation kinetics
The effect of temperature on the stability of bentonite colloids at pH≈9 and amidine latex
particles dispersed in NaCl or NaClO4 solutions (pH=5.4) was investigated. The second
order kinetics for bentonite particles at 20 and 80ºC are shown in Figure 17.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (min)
1/C
R (
Mcp
s
-1
) .
Figure 17: Inverse of the count rate as a function of time at 20C () and 80C () for bentonite
suspensions in 10-3
M NaCl or NaClO4 solutions at pH 9.
The second order slopes were the same for 10-3
M NaCl and NaClO4. At 20ºC the second
order slopes for both electrolytes were (10.0±0.1) × 10 -5
Mcps-1
min-1
and at 80ºC the
second order slopes were (2.7±0.5) × 10 -5
Mcps-1
min-1
. The faster aggregation kinetics
at 20ºC indicates lower stability at 20ºC compared with that at 80ºC.
However, the effect of temperature on the stability of amidine latex particles was the
opposite of that for bentonite. Increasing temperature had a destabilizing effect on
38
amidine latex colloids dispersed in 10-3
M NaCl or NaClO4 solutions, as can be seen in
Figure 18.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 20 40 60 80 100 120 140
Time (min)
1/C
R (
Mcp
s
-1)
.
Figure 18: Sedimentation kinetics for amidine latex particles in 10
-3 M NaCl or NaClO4
electrolyte at pH 5.7, 20C () and pH 5.4, 80C ().
Suspensions at 20C were more stable as shown by the second order slope (1.6±0.2∙10 -5
Mcps-1
min-1
), which was lower than the slope at 80C (3.7 ±0.5∙10 -5
Mcps-1
min-1
).
The zeta potential of bentonite particles at pH≈9 was independent of temperature and
stable over time as can be seen in Figure 19.
-60
-55
-50
-45
-40
-35
-30
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (min)
Zeta
po
ten
tial
(mV
)
.
Figure 19: Zeta potential of bentonite colloids in 10
-3 M NaCl or NaClO4 electrolyte at pH 9 as a
function of time at 0C (), 20C (), and 80C ().
No differences in the zeta potential were observed between NaCl and NaClO4. Table 8
shows that that the charge of the positive counter-ions strongly determined the zeta
39
potential of the particles. Since the cation in both electrolytes was monovalent Na+, the
zeta potential of bentonite particles should be the same. If the bulk composition is not
changed, the Stern layer does not change and neither does the zeta potential. Therefore,
the zeta potential should be constant with time.
Increasing temperature had a dramatic effect on the zeta potential of amidine latex
particles, as can be seen in Figure 20.
0
10
20
30
40
50
60
70
80
90
0 1000 2000 3000 4000 5000 6000 7000 8000
Time (min)
Zeta
po
ten
tial
(mV
)
.
Figure 20: Zeta potential of amidine latex colloids in 10
-3 M NaCl or NaClO4 electrolyte as a
function of time at pH 5.7, 20C (), and pH 5.4, 80C ().
Amidine latex particles are positively charged, and the Stern layer consists of the
negative counter-ions of the electrolyte in solution. No significant differences in zeta
potential were observed for the anions Cl- and ClO4
-, since both anions are monovalent.
The small influence that the different size of the anions could have on the zeta potentials
measurements, is probably of the same magnitude as the error.
Increasing temperature induces the deprotonation of amidine head groups NH2+(NH2).
75
Therefore, the zeta potential was less positive at the higher temperature (Figure 20). This
is in agreement with the fact that the pH of the suspensions at 80°C was slightly more
acidic than the pH of the suspensions at 20°C. Note that the zeta potential of amidine
particles at 80°C slightly decreased over time, indicating progressive deprotonation of the
head groups. The lower zeta potential at 80°C explains the faster aggregation kinetics
observed experimentally (Figure 18).
When the DLVO theory, equations (9)-(14), was used to analyze the effect of
temperature on the stability of bentonite particles, it showed that the repulsive energy,
and in consequence the total energy, increased with increasing temperature, as can be
seen in Figure 21.
40
-8E-18
-6E-18
-4E-18
-2E-18
0
2E-18
4E-18
6E-18
8E-18
0 1 2 3 4 5 6 7 8 9 10
Distance (nm)
VR,
VA,
VT (
J)
VA
VT
VR
Figure 21: Calculated energy curves of interaction (in J) as a function of interparticle distance
and temperature at pH 9 for bentonite particles in 10-3
M of 1:1 electrolyte. The arrows indicate
the evolution of the potential functions with increasing temperature.
These results are in agreement with the aggregation kinetics, where increasing
temperature was shown to have a stabilizing effect on bentonite suspensions. As the
maximum of the total energy increases with increasing temperature, a smaller fraction of
particles overcomes the energy barrier and the aggregation process becomes slower.
However, when the effect of temperature on the total energy was investigated for amidine
latex, the calculations showed that the total energy decreases with increasing temperature
(Figure 22).
41
-8E-18
-6E-18
-4E-18
-2E-18
0
2E-18
4E-18
6E-18
8E-18
0 1 2 3 4 5 6 7 8 9 10
Distance (nm)
VR,
VA,
VT (
J)
VR 20°C
VT 20°C
VR 80°C
VT 80°C
VA
Figure 22: Total energy curves of interaction (in J) as a function of distance and temperature for
amidine latex particles in 10-3
M of 1:1 electrolyte.
The DLVO calculations explained the faster aggregation kinetics of amidine latex
suspensions at 80°C compared with 20°C. The reduction in the surface charge by
deprotonation of the amidine groups at 80°C drastically reduced the repulsion of the
particles. The energy barrier for aggregation was lower at the highest temperature and the
particles aggregated faster.
4.2.4. Ionic strength, pH and temperature effects on the stability of
montmorillonite colloids
Since ionic strength, temperature and surface potential were found to affect the total
energy and considering that the surface charge of montmorillonite edge groups is pH
dependent, the study of the effects of temperature and ionic strength on the stability of
montmorillonite colloids was extended to other pH values.
The aggregation kinetics for montmorillonite suspensions were studied in the pH range 3
to 11. The temperatures investigated were 4, 22 and 70°C and the ionic strength was
varied from 0.002 to 0.0068 M NaCl. The second order slopes obtained are presented in
Table 10.
42
Table 10. Second order slopes for montmorillonite aggregation kinetics expressed in Mcps-1
min-1
From Table 10 it can be seen that:
- Increasing ionic strength accelerated the aggregation kinetics of montmorillonite
particles for all the temperatures and pH values studied. This was due to a decrease in the
effective repulsion between the charged colloidal particles with increasing ionic strength.
- The aggregation rate constant decreased with increasing pH. This effect became more
pronounced at higher ionic strengths and higher temperatures but was not observed at
4ºC.
- The effect of temperature was pH- and ionic strength-dependent. As the temperature
increased, the suspensions at pH≤4 were dramatically destabilized, regardless of ionic
strength. At pH≥10 the aggregation rate constant decreased with increasing temperature.
In the intermediate pH interval, the effect of increasing temperature depended on the
ionic strength.
These experimental observations can be explained by the DLVO theory. The DLVO
theory was used to calculate the maximum of the total energy for the different
temperatures, pH and ionic strength used in the experiments. The relationship is shown in
Figure 23.
Ionic
Strength
(M)
T
(°C)
pH 3
(Mcps-1
min-1
) x 105
pH 4
(Mcps-1
min-1
) x 105
pH 6
(Mcps-1
min-1
) x 105
pH 8.3
(Mcps-1
min-1
) x 105
pH 10
(Mcps-1
min-1
) x 105
pH 11
(Mcps-1
min-1
) x 105
0.002 4 7 ± 1 7.3 ± 0.9 7 ± 1 8 ± 1 9.1 ± 0.8 12 ± 1
0.002 22 10 ± 3 9 ± 2 11 ± 2 11 ± 2 12 ± 2 18 ± 4
0.002 70 280 ± 60 60 ± 30 3.7 ± 0.9 2.8 ± 0.6 2.7 ± 0.4 11 ± 3
0.003 4 9.9 ± 0.7 11.0 ± 0.7 14.2 ± 0.9 12.2 ± 0.9 12 ± 1 20 ± 2
0.003 22 130 ± 20 17 ± 1 14 ± 4 13 ± 3 16 ± 1 24 ± 3
0.003 70 600 ± 100 250 ± 20 13 ± 2 12 ± 2 14 ± 3 19 ± 4
0.005 4 15 ± 5 12 ± 3 19 ± 2 13 ± 2 13 ± 5 ---
0.005 22 290 ± 60 38 ± 5 21 ± 2 20 ± 2 27 ± 2 ---
0.005 70 600 600 16 ± 1 16 ± 3 19 ± 4 ---
0.0068 4 24 ± 2 22 ± 5 27 ± 5 24.6 ± 0.6 25 ± 5 31 ± 5
0.0068 22 600 43 ± 4 33 ± 2 28 ± 5 34 ± 6 40 ±3
0.0068 70 600 600 600 500 ± 200 22 ± 1 26 ± 3
43
0
-100
-200
-300
-400 0
0.002
0.004
0.006
0.008
-2
0
2
4
6
8
x 10-18
Ionic strength (M)Surface potential (mV)
VT
max
(J)
0
1
2
3
4
5
6
7
x 10-18
4°C
22°C
70°C
Figure 23: Maxima of the total interaction energy as a function of temperature, pH and ionic
strength.
Ionic strength determines the thickness of the double layer, equation (9). As the ionic
strength increases, the thickness of the double layer decreases and the repulsion between
particles declines. This can be related to the observed faster aggregation with increasing
ionic strength (Paper II).
Decreasing the pH induces protonation of the surface groups.76,77,78
The γ parameter
accounts for the surface potential in the repulsive energy expression in accordance with
equation (12). Since the repulsive energy is directly proportional to the square of γ, a
decrease in the surface potential reduces the repulsive energy and thus the particles
aggregate faster.
Increasing temperature increases the collision frequency and the kinetic energy of the
colloidal particles. In general, this leads to faster aggregation as observed at pH≤4 (Table
10). However, the temperature also affects the repulsion between the particles in two
opposing ways. In Figure 24, the difference between the calculated maxima of the total
energy functions at 22 and 70ºC, i.e. (Vmax70
-Vmax22
), is represented as a function of the
absolute value of the surface potential ψo.
44
-2,0E-19
0 100 200 300 400 500 600
Vm
ax
70-V
max
22
(J)
.
Surface Potential (mV)
0.002 M
0.0068 M
3.0E-19
-1.0E-19
0.0
2.0E-19
Figure 24: Difference between the calculated total energy maxima at 70 and 22ºC as a function
of the surface potential for two ionic strengths, 0.002 M and 0.0068 M NaCl.
In Figure 24 it can be seen that the direction of the temperature effect depended on the
surface potential and that the magnitude of the effect depended on the ionic strength. At
low surface potentials the repulsion decreased with increasing temperature while at high
surface potentials the repulsion increased with increasing temperature.
The increase in the repulsion with increasing temperature counteracted the temperature-
induced increase in collision frequency. Therefore a stabilization effect was observed
with increasing temperature in the suspensions in the intermediate and high pH region
(Table 10).
The parameter γ decreases with increasing temperature, equation (12). This effect alone
would lead to weaker repulsion at higher temperature. However, the magnitude of the
temperature effect on γ decreases dramatically with increasing surface potential. The
ionic medium contribution to the repulsion (ionic strength-dependent terms) is also
temperature-dependent. Contrary to γ, the medium contribution increases with increasing
temperature and decreases with increasing ionic strength. The temperature effect on the
medium contribution counteracts the effect on the surface charge contribution (γ). Hence,
for low surface potentials (low pH) the temperature effect can be mainly attributed to the
effect on the surface charge contribution (γ) while for higher potentials (high pH) the
temperature effect can be attributed to the ionic medium contribution to the repulsion.
Concerning the second question in Figure 2 referring tocolloid stability, colloid
aggregation is affected by a number of different parameters. In the context of
groundwater, ionic strength is the most important since it has the strongest impact and
shows greater variability in fracture formations than pH and temperature. Temperature,
however becomes relevant in the near field of deep geological repositories.
45
4.3. Adsorption of colloids on mineral surfaces
In analogy with colloid-colloid interactions, colloidal particles may also have the ability
to interact with the fracture surface minerals. One way to evaluate the extent of colloid
sorption is to determine distribution coefficients (Kd) between solution and minerals by
batch sorption experiments.
The conventional way to express Kd is based on the sorbent mass but a more adequate
and general way of expressing the Kd value is by normalizing it to the exposed surface
area as given in expression (29). This allows for more relevant comparison between
sorption capacities of different solids independent of particle size.
Based on measurements of the fraction of sorbate remaining in solution as a function of
the amount of solid substrate (sorbent), a linear form of equation (24) can be employed.
Kd is obtained from the slope when plotting f
fi
C
CC against
V
sm a.
The Kd values of fluorescent polystyrene latex colloids with diameters 50, 100, and 200
nm diameter, were determined in 30 mL colloidal suspensions in contact with quartz,
granite, and epidote, biotite and calcite. The values are summarized in Table 11. For
calcite, only Kd values at pH 8.5, were determined since the mineral dissolves at lower
pH.79
The sorption of latex colloids is strongly dependent on the buffer concentration and pH of
the medium as well as the type of mineral as can be seen in Table 11. These observations
suggest that sorption of colloids on mineral surfaces is governed by electrostatic forces.
Similar results were obtained for gold colloid sorption on granite and latex colloid
sorption on different minerals in the Grimsel granodiorite matrix.80,81,82
The data presented in Table 11 does not indicate any effect of particle size on colloid
sorption, rather that sorption appears to depend on colloid surface charge. The zeta
potential of 50 nm latex particles is significantly lower (less negative) than the
corresponding values for 100 and 200 nm particles (Table 2), which implies that the
electrostatic barrier between 50 nm latex colloids and mineral surfaces should be lower.
This explains the higher Kd values for 50 nm particles. The electrostatic barrier further
reduces at high ionic strength and low pH, resulting in higher Kd values for all particle
sizes, according to DLVO expressions (9-14).
Among all mineral studied, biotite and calcite show markedly higher Kd values compared
to quartz, granite, and epidote. This trend can be related to the trend in point of zero
charge (pzc) of the minerals (Table 3). A mineral showing a higher pzc, has more
positive surface charges in the pH range studied, which favours sorption of negatively
charged latex particles.
46
Table 11. Distribution coefficients (Kd) for latex colloidal particles.
Mineral,
pH
Coll size
(nm)
Kd 0.001M
(m)x106
Kd 0.01M
(m)x106
Mineral,
pH
Coll size
(nm)
Kd 0.001M
(m)x106
Kd 0.01M
(m)x106
Quartz, 8.5 50 1.2 ± 0.4 4.0 ± 0.6 Epidote, 8.5 50 1.8 ± 0.2 6 ± 1
100 0.8 ± 0.1 3.0 ± 0.7 100 0.38 ± 0.07 1.2 ± 0.2
200 0.5 ± 0.1 1.0 ± 0.2 200 0.5 ± 0.1 1.5 ± 0.2
Quartz, 6 50 1.2 ± 0.2 16 ± 2 Epidote, 6 50 1.7 ± 0.2 7 ± 1
100 0.4 ± 0.1 11 ± 3 100 0.33 ± 0.06 5 ± 2
200 0.55 ± 0.05 12 ± 1 200 0.5 ± 0.1 4 ± 1
Quartz, 4.5 50 1.7 ± 0.4 17 ± 3 Epidote, 4.5 50 2.4 ± 0.2 10 ± 2
100 1.7 ± 0.1 20 ± 4 100 0.9 ± 0.2 8 ± 2
200 1.3 ± 0.3 20 ± 3 200 0.8 ± 0.1 9 ± 2
Granite, 8.5 50 0.4 ± 0.1 3 ± 1 Biotite, 8.5 50 23 ± 2 94 ± 10
100 0.40 ± 0.04 1.1 ± 0.2 100 9 ± 2 33 ± 3
200 0.3 ± 0.1 0.92 ± 0.2 200 17 ± 3 24 ± 5
Granite, 6 50 0.8 ± 0.1 13 ± 2 Biotite, 6 50 60 ± 20 270 ± 80
100 0.7 ± 0.1 11 ± 1 100 22 ± 2 280 ± 80
200 0.7 ± 0.1 6 ± 1 200 23 ± 5 340 ± 80
Granite, 4.5 50 0.7 ± 0.1 11.1 ± 0.7 Biotite, 4.5 50 61 ± 8 260 ± 10
100 0.3 ± 0.1 20 ± 2 100 60 ± 10 330 ± 50
200 0.9 ± 0.2 12.3 ± 0.2 200 29 ± 4 300 ± 50
Calcite, 8.5 50 25 ± 6 37 ± 5
100 7 ± 1 22 ± 6
200 5.1 ± 0.5 10 ± 1
The Kd values for sorption of latex colloids on granite at 0.001 M given in Table 11 and
gold colloids reported by Alonso et al., 200980
are comparable. The similarity of the Kd
values of the two types of colloids may be attributed to the fact that the zeta potential of
gold colloids and the specific surface area of granite are quite similar in both studies.
Similar Kd values were also obtained for granite and epidote (Table 11). These minerals
are also expected to have similar pzc values, judging from the properties of the individual
minerals composing the matrix. Considering the minerals with high pzc (chlorite and
biotite), the Kd values of granite and epidote could be expected to be higher than that of
quartz. As can be seen in Table 11, this was not the case. Most likely, it is due to non-
additive effects of individual constituents on the overall surface properties of granite and
epidote. Since in biotite faces and edges have different properties, it is important which
47
parts are exposed in fracture filling minerals for the sorption. On the other hand, in
sorption experiments all surfaces are exposed.
The Kd values of Na-montmorillonite colloidal suspension in contact with quartz sand at
pH 8.5, 6, or 4.5 and buffer concentration 0.001 M are presented in Table 12.
Table 12. Distribution coefficients (Kd) for Na-montmorillonite colloidal particles in 0.001 M
buffer concentration.
Mineral pH Colloid size (nm) Kd (m) x106
Quartz 8.5 20-600 1.6 ± 0.2
Quartz 6 20-600 2.2 ± 0.2
Quartz 4.5 20-600 2.4 ± 0.2
As can be seen in Table 12, similar to latex colloids, the Kd values of Na-montmorillonite
increase with increasing pH, since the positive surface charges increases, and thus
repulsion is reduced, both for quartz and Na-montmorillonite.
Considering the process of colloid sorption onto minerals (question nr.3 in Figure 2),
colloids will be partly retained in fractures by chemical interactions.
Despite the unfavourable conditions for sorption between negative colloids and surfaces
and repulsive forces dominate between colloids and surfaces, sorption is indeed observed.
4.4. Colloid transport
So far, individual processes contributing to colloid facilitated transport have been studied.
This section deals with colloid transport in which several of the individual processes are
involved.
For all the experiments performed with polystyrene latex particles, the breakthrough
curves are symmetrical and the three colloidal sizes break simultaneously with each other
and the water front. The recovery of 50 nm colloids is significantly lower than the
recoveries of the 100 and 200 nm particles. Figure 25 shows a representative
breakthrough curves and recovery curve of transport experiments performed with latex
colloids.
48
NaCl 0.001M, Porosity 0.4, 0.03 mL/min
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Pore volumes
Ct/
Ni (m
L -1 )
.
50 nm, pH=6
100 nm, pH=6
200 nm, pH=6
50 nm, pH=8.5
100 nm, pH=8.5
200 nm, pH=8.5
NaCl 0.001M, Porosity 0.4, 0.03 mL/min
0
20
40
60
80
100
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Pore volumes
Recovery
(%
)
.E
50 nm, pH=6
100 nm, pH=6
200 nm, pH=6
50 nm, pH=8.5
100 nm, pH=8.5
200 nm, pH=8.5
HTO
Figure 25: Breakthrough and recovery curves for latex particles of 50, 100 and 200 nm diameter
in quartz transport experiments at pH 6 and 8.5 with 40% porosity and flow rate 0.03 mL min-1
.
As can be seen in Figure 25, pH has no effect on colloid transport within the limited pH
range used in the experiments. This is expected considering that, for this pH range, no
significant changes on the surface charge of quartz would take place since the pzc of
quartz is around 2, and the zeta potential of the colloidal particles is not affected
significantly according to Table 2.
Colloid recovery is always lower than 100%, and significant colloid retention takes place
in the columns. Considering that the breakthrough curves are fairly symmetric, and that
the recovery curves reach a plateau after a given time, it indicates that the immobilization
of latex particles in the column is irreversible for the time frame of the experiments. The
results from all transport experiments are summarized in Table 13.
49
Table 13. Experimental conditions and results of transport experiments.
As expected, colloid recovery increases with increasing flow rate. As the flow rate
decreases, porosity starts to have an effect on colloid recovery.
It is also interesting to note that the recovery increases with increasing particle size at the
lowest flow rate. This is not the general trend at higher flow rates.
According to the Kd values presented in Table 11, latex particles more strongly sorb to
biotite than to quartz given the lower pzc for biotite compared to quartz.
Exp. No. Flow rate
(mL min-1
)
Porosity
(%)
Column
mass
(kg)
pH/
NaCl
(M)
Colloid
Type/size
(nm)
Recovery
(%)
1 0.60±0.05 Quartz/40 0.130 8.5/
0.001
Latex 50 50±3
Latex 100 92±4
Latex 200 79±3
2 0.030±0.004 Quartz/40 0.130 8.5/
0.001
Latex 50 16±2
Latex 100 56±2
Latex 200 56±2
3 0.030±0.004 Quartz/40 0.130 6/ 0.001
Latex 50 20±2
Latex 100 56±2
Latex 200 60±2
4 0.0020±0.0002 Quartz/40 0.130 8.5/
0.001
Latex 50 15±1
Latex 100 48±2
Latex 200 50±2
5 0.60±0.05 Quartz/33 0.140 8.5/
0.001
Latex 50 52±2
Latex 100 93±4
Latex 200 69±3
6 0.030±0.004 Quartz/33 0.140 8.5/
0.001
Latex 50 35±2
Latex 100 85±4
Latex 200 71±3
7 0.0020±0.0002 Quartz/33 0.140 8.5/
0.001
Latex 50 6±1
Latex 100 31±1
Latex 200 36±1
8 0.0020±0.0002
Quartz+
Biotite
/34
0.127+
0.010
8.5/
0.001
Latex 50 2±1
Latex 100 14±1
Latex 200 24±1
9 0.0020±0.0002
Quartz+
Biotite
/35
0.124+
0.020
8.5/
0.001
Latex 50 2±1
Latex 100 9±1
Latex 200 12±1
10 0.0020±0.0002 Quartz/33 0.140 8.5/
0.001
Mont. 50-100 30±4
Mont.100-150 33±5
Mont.150-200 40±6
11 0.0020±0.0002 Quartz/33 0.140 8.5/
1-0.001
Mont. 50-100 22±3
Mont.100-150 23±3
Mont.150-200 28±4
50
In order to assess the impact of sorption on colloid retention, experiments using columns
packed with quartz mixed with two different amounts of biotite were performed at the
lowest flow rate. The resulting recovery curves are shown in Figure 26.
0.002 mL/min, 0.33-0.35 porosity
0
5
10
15
20
25
30
35
40
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Pore volumes
Recovery
(%
)
.
200 nm, quartz
200 nm, 10g biotite
200 nm, 20g biotite
Figure 26: Comparison of the recovery curves for latex colloids 200 nm diameter in transport
experiments at flow 0.002 mL min-1
using 140 g clean fine quartz and 127 or 124 g quartz mixed
with 10 and 20 g free biotite respectively.
Although changes on the pore structure may have been induced by introducing biotite in
the columns, the recovery trend appears to reflect the increasing sorption capacity of the
column. This observation and the fact that the relative recovery trends in Table 11 at the
lowest flow rates in pure quartz columns reflect the Kd values of the different particle
sizes point in the direction that sorption is of significance at low enough flows. At higher
flows however, the relationship between recovery and Kd values is less obvious,
indicating that at fast flow, sorption equilibrium is not reached. It should be noted that the
impact of physical filtration would also decrease with increasing flow.
The previous observations can be better illustrated by plotting the recovery against the
total sorption capacity (filling material mass x specific surface area x Kd) of the columns
for the different experiments as shown in Figure 27.
51
0
10
20
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
0.002 ml/min, quartz 33% porosity
0.002 ml/min, quartz + 10 g biotite, 34 % porosity0.002 ml/min, quartz + 20 g biotite, 34 % porosity0.002 ml/min, quartz 40 % porosity
0.03 ml/min, quartz 33 % porosity
0.03 ml/min, quartz 40 % porosity
Sorption capacity (m3)
Figure 27: Relationship between experimental recovery and column sorption capacity.
From Figure 27 it becomes evident that at low flow where the residence time is long, and
porosity of 33% where the available surface area is larger, sorption appears to be the
dominant retention process. In the experiments with higher porosities and/or higher flow
rates, the higher recoveries observed are indicative of less interaction between the latex
particles and the mineral surfaces. The lower recoveries of the 50 nm latex particles can
be explained by higher Kd values than for the larger latex particles. The 50 nm latex
particles possess a lower negative charge than the 100 and 200 nm particles giving higher
Kd values. It should be noted that size exclusion would give the same trend, however, the
effect would most probably be larger since the difference in the particle sizes are quite
large. The resemblance in the 100 and 200 nm particles behavior, would suggest that it is
more likely that the effect is reflecting sorption with similar sorption affinities for the
mineral.
The transport of colloidal Na-montmorillonite particles was studied at the lowest flow
and porosity in columns packed with fine quartz sand. The breakthrough and recovery
curves are shown in Figure 28.
52
Na-montmorillonite, porosity 0.33, flow 0.002 ml/min
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Pore volumes
Ct/N
i (m
L
-1 )
.
50-100 nm
100-150 nm
150-200 nm
Na-montmorillonite, porosity 0.33, flow 0.002 ml/min
0
5
10
15
20
25
30
35
40
45
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Pore volumes
Recovery
(%
)
.
50-100 nm
100-150 nm
150-200 nm
Figure 28: Breakthrough and recoveries for the Na-montmorillonite particles in three size classes
in a quartz column with porosity of 33 % and flow of 0.002 mL min-1
.
One important difference between the breakthrough curves for the Na-montmorillonite
particles compared to latex colloids is that the former are not symmetrical. The tailings in
the breakthrough curves indicate reversible retention, while the retention in the transport
of latex colloids appears to be irreversible, at least for this experimental time frame.
Another possibility is that small colloids are generated from the larger aggregates trapped
in the column. The released smaller particles are then able to be transported.
53
5. Concluding remarks
Concentration of montmorillonite colloids outside the barrier can be predicted
from the results of laboratory experiments on generation and sedimentation of
montmorillonite colloids. The concentrations obtained from the two different
types of experiments are identical. This implies that simple sedimentation
experiments can be used to determine the pseudo-equilibrium concentrations.
The stability of montmorillonite colloids decreases with increasing ionic strength
and decreasing pH as expected. The pH and ionic strength were proved to be
critical parameters influencing the effect of temperature. At intermediate pH, the
effect of increasing temperature stabilized the suspensions with ionic strength
lower than 0.0068 M. At pH≤4 increasing temperature dramatically reduced the
stability of montmorillonite suspensions independently of ionic strength. At
pH≥10 increasing temperature stabilized the suspensions for all ionic strengths
studied.
Montmorillonite suspensions were more sensitive to ionic strength changes
induced by increasing CaCl2 than by NaCl. The CCC value obtained for Ca-
montmorilonite in a solution containing CaCl2 was two orders of magnitude lower
than that for Na-montmorillonite in a solution containing NaCl.
Latex and montmorillonite colloids were shown to sorb to minerals under
unfavourable conditions. This process appears to be governed by the electrostatic
forces between minerals and colloids as reflected by mineral pzc and colloid zeta
potential. Consequently, sorption is affected by water chemistry in a similar
fashion to colloid-colloid interactions.
Colloids break through the columns packed with different fracture filling minerals
under advective regime. Colloid sorption to minerals contributes to colloid
retention during transport reflecting pzc trends of the minerals. In the present
experiments, Na-montmorillonite retention is reversible while for latex colloids
retention is irreversible in the experimental time frame.
The DLVO theory provided a qualitative explanation for effects of the different
parameters on colloid generation, stability of the suspensions and sorption.
54
6. Acknowledgements
Many people have contributed to this thesis in many different ways. Therefore I would
like to thank:
- my two supervisors Prof. Mats Jonsson and Dr. Susanna Wold for all I learned from you
these years not only during our discussions. Your interest in science increased my
motivation.
- the Swedish Nuclear Fuel and Waste Management Co., SKB, for funding this work and
in particular, Dr. Ignasi Puigdomenech and Dr. Kastriot Spahiu for all the support,
discussions and good advices.
- Dr. Claude Degueldre for giving me the opportunity to visit the Paul Scherrer Institute
(PSI) and learn more about colloids, and to all the people that I meet there.
- to all of you that have contributed with diverse practical help: everyone at Nuclear
Chemistry (KTH) and in particular Michael Holmboe, Anders Puranen and Kjell
Svärdström for support in the lab; Dr. Alan Snedden and Dr. Andreas Fischer at Inorganic
Chemistry (KTH) for assistance with SEM and XRD analysis; Ola Karnland at Clay
Technology for providing montmorillonite samples; Oskar Sigurdsson at Äspö hardrock
laboratory, Sweden for assistance during mineral sample collection; Jan Olov Nyström
and Henrik Skogby at the Royal Museum of Natural History in Stockholm for providing
biotite samples; Luc van Loon, Martin Glaus and Sabina Frick at PSI for help with
generation experiments.
- I am also grateful to Prof. Emeritus Trigve Eriksen, Prof. Ivars Neretnieks, Michael
Holmboe, Dr. Luc van Loon and Martin Glaus for illuminating discussions.
- Thanks to all the former and present workmates from the Nuclear Chemistry and
Inorganic Chemistry divisions at KTH for creating a scientific and enjoyable atmosphere
at work.
- My family and friends here and there, thank you for the faith, the good advice and the
great interest that you have always shown in my work.
- Special thanks to Victor for your facility to listen, understand and see the positive in the
difficulties.
55
7. Bibliography
[1] Atteia, O., Kozel, R., Particle size distributions in waters from a karstic aquifer: from
particles to colloids. J. of Hydrology 201 (1997) 102-119.
[2] Dearlove, J.P.L., Longworth, G., Ivanovich, M., Kim, J.I., Delakowitz, B., Zeh, P., A
study of groundwatercolloids and their geochemical interactions with natural
radionuclides in Gorleben aquifer systems. Radiochimica Acta 52/53 (1991) 83-89.
[3] Gschwend, P.M., Backhus, D.A., MacFarlane, .I.K., Page, A.L., Mobilization of
colloids in groundwater due to infiltration of water at a coal ash disposal site. J. Contam.
Hydrol. 6 (1990) 307-320.
[4] Degueldre, C., Baeyens, B., Goerlich, W., Riga, J., Verbist, J., Stadelmann, P.,
Colloids in water from a subsurface fracture in granitic rock, Grimsel Test Site,
Switzerland. Geochem. Cosmochim. Acta 53 (1989) 603 -610.
[5] Sposito, G., The surface chemistry of soils, Oxford University Press, New York, 1984.
[6] McCarthy, J.F., Zachara, J.M., Subsurface transport of contaminants: Mobile colloids
in the subsurface environment may alter the transport of contaminants. Environmental
Science and Technology, 23 (1989) 496-502.
[7] Kersting, A., Efurd, D., Finnegan, D., Rokop, D., Smith, D., Thompson, J., Migration
of plutonium in ground water at the Nevada Test Site. Nature, 397 (1999) 56-59.
[8] www.andra.fr
[9] www.nirond.be
[10] http://www.posiva.fi/en
[11] www.skb.se
[12] Final storage of spent nuclear fuel - KBS-3, I General; II Geology; III Barriers; IV
Safety; - Summary. Swedish Nuclear Fuel Supply Co/Division KBS (SKBF/KBS) 1983
[13] SR 97-post-clossure safety, SKB TR-99-06, Main Report, Vol. I., II. SKB, 1999.
[14] Long-term safety for KBS-3 repositories at Forsmark and Laxemar – a first
evaluation Main Report of the SR-Can project, SKB Technical Report, TR-06-09, 2006.
[15] Segall, R.L., Smart, R.St.C., Turner, P.S., Surface and Near- Surface Chemistry of
Oxide Materials. Materials Science Monographs 47, edited by J. Nowotny and L.-C.
Dufour, Elsevier Science Publishers B. V., Amsterdam, 1988.
56
[16] King F., Corrosion of copper in alkaline chloride environments. SKB Technical
Report, TR-02-25, 2002.
[17] Börgesson, L., Hernelind, J., Earthquake induced rock shear through a deposition
hole. Influence of shear plane inclination and location as well as buffer properties on the
damage caused to the canister. SKB Technical Report, TR-06-43, 2006.
[18] Karnland, O., Olsson, S., Nilsson, U., Sellin, P., Mineralogy and sealing properties
of various bentonites and smectite-rich clay materials. SKB Technical Report, TR-06-30,
2006.
[19] Vejmelka, P., Fanghaenel, Th., Kienzler, B., Korthaus, E., Römer, J., Schüssler, W.,
Artinger, R., Sorption and migration of radionuclides in granite (HRL Äspö, Sweden).
FZKA 6488, 2000.
[20] Andersson, J., Ström, A., Svemar, C., Almén, K-E., Ericsson, L.O., What
requirements does the KBS-3 repository make on the host rock? Geoscientific suitability
indicators and criteria for siting and site evaluation. SKB Technical Report, TR-00-12,
2000.
[21] Bergelin, A., Lindquist, A., Nilsson, K., Nilsson, A.-C., Forsmark site investigation.
Hydrochemical characterization in borehole KFM10A. Results from three investigated
borehole sections: 298.0–305.1 m, 436.9–437.9 m, 478.0–487.5 m. SKB Technical report
P-07-42 2007.
[22] Wass, E., Groundwater flow measurements in permanently installed boreholes. Test
campaign no. 4, 2008. Forsmark site investigation. SKB Technical report P-09-30, 2009.
[23] Auqué, L. F., Gimeno, M.J., Gómez, J.B., Puigdomenech, I., Smellie, J., Tullborg,
E-L., Groundwater chemistry around a repository for spent nuclear fuel over a glacial
cycle. Evaluation for SR-Can. SKB Technical Report, TR-06-31, 2006.
[24] Boulton, G.S., Caban, P.E., van Gijssel, K., Groundwater flow beneath ice sheets:
Part I — Large scale patterns. Quaternary Science Reviews, 14 (1995) 545-562.
[25] Neretnieks, I., Flow and transport through a damaged buffer- exploration of the
impact of a cemented and an eroded buffer. SKB Technical Report, TR-06-33, 2006.
[26] Neretnieks, I., Liu, J., Physical and chemical stability of the bentonite buffer, R-06-
103, SKB, 2006.
[27] Karnland, O., Sandén, T., Johannesson, L.-E., Eriksen, T. E., Jansson, M., Wold, S.,
Pedersen, K., Motamedi, M., Rosborg, B., Long term test of buffer material. Final Report
on the pilot parcels. SKB Technical Report, TR-00-22, 2000.
57
[28] Puigdomenech, I., Hydrochemical stability of groundwaters surrounding a spent
nuclear fuel repository in a 100000 year perspective, SKB Technical Report, TR-01-28,
2001.
[29] Neretnieks, I., Diffusion in the rock matrix: an important factor in radionuclide
retardation?. J. Geophys. Res. 85 B8, (1980) 4379–4397.
[30] Vilks, P., Bachinski, D.B., Colloid and suspended particle migration experiments in
a granitic fracture. Fourth International conference on the chemistry and migration of
actinides and fission products in the geosphere. Charleston, SC, USA, proceedings.
Radiochimica acta, special issue, 66/67 (1993) 229-234.
[31] Norrish, K., The swelling of montmorillonite. Discuss. Faraday Soc. 18 (1954) 120-
134.
[32] Hunter, R. J., Introduction to modern colloid science. Oxford Science Publications,
1993.
[33] Shchukin, E.D., Pertsov, A.V., Amelina E.A., Zelenev, A.S., Colloid and Surface
Chemistry. Studies in Interface Science, vol. 12, Elsevier, 2001, p.214.
[34] Shaw, D.J., Introduction to colloid and surface chemistry. Butterworths & Co
Publishers, 1970.
[35] Ran, Y., Fu, J.M., Sheng, G.Y., Beckett R., and Hart, B.T., Fractionation and
composition of colloidal and suspended particulate materials in rivers. Chemosphere,
41 (2000) 33-43.
[36] Plaschke, M., Schäfer, T., Brundschuh, T., Ngo Manh, T., Knopp, R., Geckeis H.,
.Kim, J.I., Size Characterization of Bentonite Colloids by Different Methods. Anal.
Chem., 73 (2001) 4338-4347.
[37] Cadene, A., Durand-Vidal, S., Turq, P., Brendle, J., Study of individual Na-
montmorillonite particles size, morphology, and apparent charge. J. Coll. Interface Sci.,
285 (2005) 719-730.
[38] McCarthy, J.F., Degueldre, C., Sampling and Characterization of Colloids and
Particles in Groundwater for Studying Their Role in Contaminant Transport.
Environmental Particles, Vol.2, J. Buffle and H. P. van Leeuwen (Eds.), Lewis Publisers,
Boca Raton, FL, 1993 p. 247-315.
[39] Ryan, J.N., Elimelech, M., Colloid mobilization and transport in groundwater.
Colloids Surfaces A: Physicochemical Eng. Aspects 107 (1996) 1-56.
58
[40] Karnland, O., Olsson, S., Nilsson, U., Sellin, P., Experimentally determined
swelling pressures and geochemical interactions of compacted Wyoming bentonite with
highly alkaline solutions. Physics and Chemistry of the Earth 32 (2007) 275–286.
[41] Scatchard, G., Equilibrium in solutions. Surface and Colloid Chemistry”, p.260-261,
Harvard University Press, 1976.
[42] Stechemesser, H., Dobiáš, B., Coagulation and Flocculation. 2nd
edition, Surfactant
Science Series vol. 126, CRC Press Taylor and Francis, 2005, p.81-88.
[43] J. Lyklema, Fundamentals of Interface and Colloid Science. Vol. IV, Elsevier
academic press, 2005.
[44] B.V. Derjaguin, L. Landau, Acta Physicochim URSS, 14 (1941) 633.
[45] Reerink, H., Overbeek, J.Th.G., Theory of the stability of lyophobic colloids.
Elsevier, Amsterdam 1948.
[46] Hiemenz, P.C., Principles of colloid and surface chemistry. 2nd
ed. Marcel Dekker,
inc. New York, 1986.
[47] Gillham, R.W., Cherry, J.A., Predictability of solute transport in diffusion controlled
hydrogeologic regimes, Proceedings of the Symposium on Low-Level Waste Disposal:
Facility Design, Construction and Operating Practices, 1982a.
[48] Gillham, R.W., Cherry, J.A., Contaminant migration in saturated unconsolidated
geologic deposits, Recent Trends in Hydrogeology, The Geological Society of America,
Special Paper 189, Boulder, Colorado, 31–62, 1982b.
[49] Davis, S.N., “Porosity and permeability in natural materials. Flow through Porous
Media, DeWiest, R. J. M. (ed.), 53-89. Academic Press, New York, NY, 1969.
[50] Darcy, H., Les Fontaines Publiques de la Ville de Dijon. Victor Dalmont, Paris,
1856.
[51] Hiscock, K.M., Hydrogeology: principles and practice Blackwell Pub., Malden, MA
2005.
[52] Fetter, C.W., Applied Hydrogeology. MacMillan, New York, NY, 1994.
[53] Stumm W., Morgan, J.J., Aquatic Chemistry: Chemical equilibria and rates in
natural waters. 3rd edn., Wiley, New York, 1996.
[54] Malmström, M., Banwart, S., Lewenhagen, J., Duro, L., Bruno J., The dissolution of
biotite and chlorite at 25°C in the near-neutral pH region Journal of Contaminant
Hydrology, 21 (1996) 201-213.
59
[55] Sverjensky, D.A., Zero-point-of-charge prediction of crystal chemistry and salvation
theory, Geochimica et Cosmochimica Acta 58 (1994), pp. 3123–3129.
[56] Sondi, I., Bian, J., Pravdi, V., Electrokinetics of Pure Clay Minerals Revisited. J.
Coll. Interface Sci., 178 (1996) 514-522.
[57] Rossé, P., Loizeau, J.L., Use of single particle counters for the determination of the
number and size distributions of colloids in natural surface waters. Colloids and Surfaces
A: Physicochem. Eng. Aspects 217 (2003) 109-120.
[58] Liquid sampler-50 operator´s manual, P/N M10205 Rev A, Particle measuring
systems, Inc. 2004
[59] Ledin, A., Karlsson, S., Düker, A., Allard, B., Applicability of photon correlation
spectroscopy for measurement of concentration and size distribution of colloids in
natural waters. Anal. Chim. Acta, 281 (1993) 421-428.
[60] Xu, R., Particle Characterization: Light Scattering Method, Hingham, MA, USA:
Kluwer Academic Publishers, 2000. p 223-269.
[61] Miller, W.M., Alexander, W.R., Chapman, N.A., McKinley, I.G., Smellie, J.A.T.,
Geological Diposal of Radioactive Wastes and Natural Analogues. Waste Management
Series, vol. 2, Pergamon, Amsterdam, 2000.
[62] Vold, R.D., Vold, M.J., Colloid and Interface Chemistry. Addison Wesley
Publishing Company, 1983
[63] Skoog, D.A., Leary, J.J., Principles of instrumental analysis, Saunders Golden
Sunburst Series, 1992.
[64]http://www.varianinc.com/image/vimage/docs/products/spectr/fluoro/brochure/1757.
[65] Choppin, G.R.R., Rydberg, J., Liljenzin, J.-O., Radiochemistry & nuclear chemistry
2ed., Elsevier Science & Technology Books, 1995.
[66] Missana, T., Alonso, U., Turrero, M.J., Generation and stability of bentonite
colloids at the bentonite/granite interface of a deep geological radioactive waste
repository. J. of Contam. Hydrology, 61 (2003) 17-31.
[67] Bessho, K., Degueldre, C., Generation and sedimentation of colloidal bentonite
particles in water. App. Clay Sci., 43 (2009) 253-259.
[68] Pusch, R.. Experimental study of the effect of high porewater salinity on the physical
properties of a natural smectitic clay. SKB Technical Report, TR-01-07, (2001).
60
[69] Tang, L., Sparks, D.L., Cation-exchange kinetics on montmorillonite using
pressure-jump relaxation. Soil Sci. Soc. Am. J., 57 (1993) 42- 46.
,
[70] Frey E. and Lagaly, G., Selective coagulation in mixed colloidal suspensions. J. Coll
Interface Sci., 70 (1978) 1.
[71] Tombácz, E., Balázs, J., Lakatos J., and Szántó, F., Influence of the exchangeable
cations on stability and rheological properties of montmorillonite suspensions. Colloid
Polym. Sci., 267 (1989) 1016-1025
[72] Chheda, P., Grasso D., and van Oss, C.J., Impact of ozone on stability of
montmorillonite suspensions. J. Coll. Interface Sci., 153 1 (1992) 226-236.
[73] Lagaly G., Ziesmer, S., Colloid chemistry of clay minerals: the coagulation of
montmorillonite dispersions. Advances Coll. Interface Sci., 100-102 (2003) 105-128.
[74] SKB, 2004a. Interim initial state report for the safety assessment SR-Can. SKB
Report R-04-35, 2004.
[75] Homola, A., James, R.O., Preparation and characterization of amphoteric
polystyrene lattices. J. Coll. Interface Sci., 59 (1977) 123-34.
[76] Avena, M.J., De Pauli, C.P., Proton Adsorption and Electrokinetics of an
Argentinean Montmorillonite. J. Coll. Interface Sci., 202 (1998) 195-204.
[77] Kraepiel, A.M.L., Keller, K., Morel, F.M.M., A Model for Metal Adsorption on
Montmorillonite., J. Coll. Interface Sci., 210 (1999) 43-54.
[78] Tombácz. E., Ábrahám, I., Gilde, M., Szántó, F., The pH-dependent colloidal
stability of aqueous montmorillonite suspensions. Coll. and Surfaces, 49 (1990) 71-80.
[79] Compton, R.G., Pritchard, K.L., The Dissolution of Calcite at pH > 7: Kinetics and
Mechanism, Phil. Trans. R. Soc. Lond. A 330 (1990) 47-70.
[80] Alonso, U., Missana, T., Patelli, A., Ceccato, D., Albarran, N., García-Gutiérrez, M.
Lopez-Torrubia, T., Regato, V., Quantification of Au nanoparticle retention on a
heterogeneous rock surface. Colloids Surf., A. 347: 1-3, (2009) 230-238.
[81] Taboada-Serrano, P., Chin, C.-J., Yiacoumi, S., Tsouris, C., Modeling aggregation
of colloidal particles. Current Opinion Coll. Interface Sci., 10 (2005) 123-132.
[82] Filby, A., Plaschke, M., Geckeis, H., Fanghänel, Th.,. Interaction of latex colloids
with mineral surfaces and Grimsel granodiorite. J.Contam. Hydrol., 102 (2008) 273-284.